CN116719927A

CN116719927A - Digest generation method, digest generation device, intelligent terminal and computer-readable storage medium

Info

Publication number: CN116719927A
Application number: CN202310492894.XA
Authority: CN
Inventors: 高建清; 王玮; 苏文畅; 戚婷; 江嘉辉; 孔常青; 闫莉; 徐迪迪; 熊世富; 潘嘉; 刘聪; 胡国平; 魏思; 王士进; 刘权
Original assignee: iFlytek Co Ltd
Current assignee: iFlytek Co Ltd
Priority date: 2023-04-06
Filing date: 2023-04-28
Publication date: 2023-09-08

Abstract

The application discloses a method, a device, an intelligent terminal and a computer readable storage medium for generating a abstract, wherein the method comprises the following steps: acquiring a selected text selected by a user from the texts to be processed; obtaining abstract text at least based on the selected text; and displaying the abstract text. By the scheme, the convenience in acquiring the abstract text can be improved.

Description

Digest generation method, digest generation device, intelligent terminal and computer-readable storage medium

Technical Field

The present application relates to the field of data processing technologies, and in particular, to a method and apparatus for generating a digest, an intelligent terminal, and a computer readable storage medium.

Background

With the advent of the data age, users can acquire a large amount of text data in various life, work, study and other scenes, so that users increasingly expect to be able to extract abstract text from the text data conveniently and reliably. In the prior art, text data is usually only displayed for a user, and the process of acquiring abstract text based on the text data relies on the user to review the text data and extract the required abstract text from the text data, which is very inconvenient for the user. In view of this, how to improve the convenience of obtaining the abstract text is a problem to be solved.

Disclosure of Invention

The application mainly solves the technical problem of providing a method, a device, an intelligent terminal and a computer readable storage medium for generating the abstract, which can improve the convenience of acquiring the abstract text.

In order to solve the above technical problems, a first aspect of the present application provides a method for generating a summary, including: acquiring a selected text selected by a user from the texts to be processed; obtaining abstract text at least based on the selected text; and displaying the abstract text.

In order to solve the above technical problem, a second aspect of the present application provides a summary generating apparatus, including: the system comprises an acquisition module, a generation module and a display module, wherein the acquisition module is used for acquiring selected texts selected by a user from texts to be processed; the generation module is used for obtaining abstract text at least based on the selected text; and the display module is used for displaying the abstract text.

In order to solve the technical problem, a third aspect of the present application provides an intelligent terminal, which includes a display screen, a memory and a processor, wherein the display screen and the memory are respectively coupled to the processor, the display screen is at least used for displaying content to a user and selecting content for the user, the memory stores program instructions, and the processor is used for executing the program instructions to implement the summary generation method described in the first aspect.

To solve the above-mentioned technical problem, a fourth aspect of the present application provides a computer-readable storage medium having stored thereon program instructions which, when executed by a processor, implement the digest generation method described in the first aspect.

According to the scheme, after the user selects at least part of texts in the texts to be processed, the texts selected by the user in the texts to be processed are obtained to serve as the selected texts, the abstract texts corresponding to the selected texts are generated at least based on the selected texts, so that the abstract texts are at least related to the selected texts, namely, the reference range in abstract text generation at least comprises the selected texts, and after the abstract texts are obtained, the obtained abstract texts are displayed. Therefore, the user only needs to check the text and select a part of the text, the process of selecting the selected text by the user can meet the requirement of free selection of the user, and when the user selects at least a part of the text in the text to be processed as the selected text, the abstract text with high matching degree with the selected text can be obtained, and the convenience of the user for obtaining the abstract text is improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present application, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art. Wherein:

FIG. 1 is a schematic flow chart of an embodiment of a summary generation method of the present application;

FIG. 2 is a schematic flow chart of another embodiment of the summary generation method of the present application;

fig. 3 is a schematic view of an application scenario of an embodiment corresponding to step S201 in fig. 2;

fig. 4 is a schematic view of an application scenario of another embodiment corresponding to step S201 in fig. 2;

fig. 5 is a schematic view of an application scenario of a further embodiment corresponding to step S201 in fig. 2;

FIG. 6 is a schematic flow chart diagram of a further embodiment of the summary generation method of the present application;

fig. 7 is a schematic diagram of an application scenario of an embodiment corresponding to step S602 in fig. 6;

FIG. 8 is a flow chart of a further embodiment of the summary generation method of the present application;

fig. 9 is a schematic diagram of an application scenario of an embodiment corresponding to step S802 in fig. 8;

fig. 10 is a schematic diagram of an application scenario of an embodiment corresponding to step S806 in fig. 8;

fig. 11 is a schematic view of an application scenario of another embodiment corresponding to step S806 in fig. 8;

FIG. 12 is a schematic diagram showing the structure of a summary generating apparatus according to an embodiment of the present application

FIG. 13 is a schematic structural diagram of an embodiment of a smart terminal according to the present application;

fig. 14 is a schematic diagram of a computer-readable storage medium according to an embodiment of the present application.

Detailed Description

The following description of the embodiments of the present application will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.

The terms "system" and "network" are often used interchangeably herein. The term "and/or" is herein merely an association relationship describing an associated object, meaning that there may be three relationships, e.g., a and/or B, may represent: a exists alone, A and B exist together, and B exists alone. In addition, the character "/" herein generally indicates that the front and rear associated objects are an "or" relationship. Further, "a plurality" herein means two or more than two.

The abstract generation method provided by the application is realized by depending on the application on the intelligent terminal or the intelligent terminal at least integrated with a text processing function, and an execution subject corresponding to the abstract generation method provided by the application is a processor of the intelligent terminal.

Referring to fig. 1, fig. 1 is a flow chart of an embodiment of a summary generating method according to the present application, the method includes:

s101: and acquiring a selected text selected by the user from the texts to be processed.

Specifically, after the user selects at least part of the text in the text to be processed, the text selected by the user in the text to be processed is obtained as the selected text.

Further, the user can select the text to be processed in a preset circle selection mode, and after the user selects at least part of the text from the text to be processed in the preset circle selection mode, the selected text selected by the user in the preset circle selection mode is obtained. The preset circle selection mode comprises at least one of drawing operation, hooking operation and dragging operation.

In an application mode, responding to the detection of the drawing operation of a user, and acquiring a text selected by the user from the texts to be processed through the drawing operation as a selected text. The drawing operation includes drawing a closed figure or drawing a preset symbol, and the application does not limit the closed figure and the preset symbol.

In an application scene, a circle selection description of a drawing operation is presented, so that a user can learn that a graph (such as an ellipse or a polygon) closed from beginning to end is drawn, texts included in the graph can be used as selected texts, and in response to the user selecting part of texts in the texts to be processed by using the drawing operation, the texts selected by the user in the texts to be processed by the drawing operation are acquired as the selected texts.

In another application scenario, a circle selection description of the drawing operation is displayed, so that a user can learn that a text corresponding to a preset symbol (such as brackets or straight lines) is taken as a selected text, a text in the brackets is taken as a selected text before and after the preset symbol, or a text corresponding to the straight lines is taken as a selected text, and in response to the user selecting a part of texts in the texts to be processed by the drawing operation, the text selected by the user in the texts to be processed by the drawing operation is taken as the selected text.

In another application mode, in response to detection of a checking operation of a user, a text selected by the user from the texts to be processed through the checking operation is obtained as a selected text. The text to be processed comprises a plurality of first candidate items, the choosing operation comprises at least one of clicking, double clicking and sliding, and the shape of the first candidate items and the triggering form of the choosing operation are not particularly limited.

In an application scene, setting and displaying first candidate items by utilizing a rectangular frame before each natural paragraph in the text to be processed, and acquiring the text selected by the user in the text to be processed through the checking operation as the selected text after the user checks at least part of the first candidate items through the checking operation and confirms the first candidate items.

In another application scene, setting and displaying at least part of first candidates in the text to be processed by using circles, wherein the first candidates are related to the number of characters of the natural paragraph, the number of characters corresponding to each first candidate exceeds a threshold value of the number of characters, and acquiring the text selected in the text to be processed by the user through the checking operation as the selected text after the user checks at least part of the first candidates through the checking operation and confirms the first candidates.

In another application mode, in response to detection of a drag-and-drop operation of a user, a text selected by the user from the texts to be processed through the drag-and-drop operation is obtained as a selected text. The drag operation includes continuous motion operation after clicking or touching, and the mode of the drag operation is not particularly limited in the application.

In an application scene, a drag-and-drop description of drag-and-drop operation is displayed, so that a user can learn that the dragged and dropped text can be used as a selected text through the cursor drag-and-drop operation, and the user responds to the drag-and-drop operation to select a part of text in the text to be processed, so that the text selected by the user in the text to be processed through the drag-and-drop operation is obtained as the selected text.

In another application scene, a drag-and-drop description of drag-and-drop operation is displayed, so that a user can learn that the dragged and dropped text can be used as a selected text through touch drag-and-drop, and the text selected by the user in the text to be processed through the drag-and-drop operation is obtained as the selected text in response to the fact that the user selects a part of the text in the text to be processed through the drag-and-drop operation.

In a specific application scenario, a text to be processed is displayed on a touch screen of an intelligent terminal (such as a mobile phone or a tablet personal computer), an operation of drawing an ellipse or drawing brackets on the touch screen by a user is detected, and a text included in the ellipse is used as a selected text or a text between a front bracket and a rear bracket is used as a selected text.

In another specific application scenario, a text to be processed is displayed on a display screen of an intelligent terminal (such as a portable computer or a desktop computer), and a check box and a plurality of single-selection boxes are arranged in the text to be processed, wherein after the check boxes are checked, all the single-selection boxes are checked, and a user is detected to use texts corresponding to the checked single-selection boxes as checked texts through operations of checking the check boxes or at least part of the single-selection boxes.

In another specific application scenario, the text to be processed is displayed on a display screen of an intelligent terminal (such as a portable computer or a desktop computer), the text selected by the user through cursor dragging in the text to be processed is detected, and the text selected when the user drags the cursor is used as the selected text.

S102: based at least on the selected text, a summary text is obtained.

Specifically, extracting key content from selected texts to obtain abstract texts, or extracting key content from at least part of selected texts and unselected texts to obtain abstract texts, wherein the unselected texts comprise texts except the selected texts in the texts to be processed.

In an application mode, a first keyword and a first key sentence in the selected text are extracted, the first keyword and the first key sentence are integrated, and a abstract text corresponding to the selected text is obtained, so that the requirements of a user selected by the user are matched, and the convenience of the user for obtaining the abstract text is improved.

In another application mode, a first keyword and a first key sentence in the selected text are extracted, a reference text matched with the first keyword and the first key sentence is obtained from the unselected text based on the first keyword and the first key sentence, a second keyword and a second key sentence in the reference text are extracted, and the first keyword and the first key sentence, the second keyword and the second key sentence are integrated to obtain a summary text corresponding to the selected text, so that the summary text has higher key content conformity with the selected text, and convenience of a user for obtaining the summary text and accuracy of the obtained summary text are improved.

In another application mode, semantic information corresponding to the selected text is obtained, a text matched with the semantic information of the selected text is extracted from the unselected text and is used as a reference text, a first keyword and a first key sentence in the selected text and a second keyword and a second key sentence in the reference text are extracted, and the first keyword and the first key sentence, the second keyword and the second key sentence are integrated to obtain a abstract text corresponding to the selected text, so that the abstract text has higher semantic conformity with the selected text, and the convenience of obtaining the abstract text and the accuracy of obtaining the abstract text by a user are improved.

In an application scene, a first keyword and a first key sentence in the selected text are extracted based on the semantics of the selected text, and the first keyword and the first key sentence are integrated to obtain a abstract text corresponding to the selected text.

In another application scene, extracting a first keyword and a first key sentence in a selected text, extracting a reference text from the unselected text based on the frequency of outgoing lines of the first keyword and the first key sentence in the unselected text, wherein the frequency of occurrence of the first keyword or the first key sentence in the reference text is greater than a frequency threshold, the frequency threshold can be any positive integer (such as 2, 3 or 4), and the like, extracting a second keyword and a second key sentence in the reference text, and integrating the first keyword and the first key sentence, the second keyword and the second key sentence to obtain a abstract text corresponding to the selected text.

In a further application scene, the selected text and the unselected text are sent to a pre-trained semantic extraction model, the output of the semantic extraction model is obtained, semantic information corresponding to the selected text and semantic information corresponding to the unselected text are obtained, content matched with the semantic information of the selected text is extracted from the unselected text, a reference text is obtained, a first keyword and a first key sentence in the selected text and a second keyword and a second key sentence in the reference text are extracted, and the first keyword and the first key sentence and the second keyword and the second key sentence are integrated, so that a abstract text corresponding to the selected text is obtained.

S103: and displaying the abstract text.

Specifically, after the abstract text is obtained, the obtained abstract text is displayed. Therefore, the user can obtain the abstract text with high matching degree with the selected text only by selecting and selecting part of the texts perceived by the user, and the convenience of the user for obtaining the abstract text is improved.

In an application mode, a display area matched with the abstract text is obtained, the abstract text is displayed in the display area, and when the user is detected to select an editing option of the display area, the display area is adjusted to be an editable area.

In another application, the abstract text is presented using the editable box until the user selects the confirmation option of the editable box, which is adjusted to the abstract presentation area. The user can add, delete and modify the abstract text in the editable box.

In an application scene, a display area matched with the abstract text is obtained in a display interface of the text to be processed, the abstract text is displayed in the text to be processed, after an editing option of the display area selected by a user is detected, the display area is adjusted to be an editable area, and after a confirmation option selected by the user is detected, the abstract text in the display area is adjusted to be independent of other interfaces of the display interface for storage.

In another application scene, other interfaces which are not overlapped with the display interface of the text to be processed are obtained, display areas matched with the abstract text are obtained in the other interfaces, the abstract text is displayed by utilizing the editable frame, so that a user can perform addition, deletion and modification operations in the editable frame until the user selects a confirmation option of the editable frame, the editable frame is adjusted to be the abstract display area, and the abstract text in the abstract display area is stored in the corresponding interface.

Referring to fig. 2, fig. 2 is a flow chart of another embodiment of the summary generating method according to the present application, the method includes:

s201: and acquiring a selected text selected by the user from the texts to be processed.

At least one first candidate item is displayed in the text to be processed, wherein each first candidate item corresponds to one text block.

In an implementation scenario, before obtaining the selected text selected by the user among the texts to be processed, the method further includes: displaying at least one first candidate item in the text to be processed; wherein each first candidate corresponds to a block of text.

Specifically, referring to fig. 3, fig. 3 is a schematic view of an application scenario corresponding to step S201 in fig. 2, at least one first candidate item (e.g., a rectangular box in fig. 3) is set in the text to be processed, the first candidate item is displayed in the text to be processed, and each first candidate item corresponds to a text block. Wherein the text block comprises at least part of the content in the text to be processed.

Further, when the user selects the first candidate (e.g., the rectangle box checked in fig. 3), the contents in the text blocks corresponding to the first candidate selected by the user are all used as the selected text, so that the convenience of text selection by the user can be improved.

In an application mode, generating at least one first candidate item based on semantics corresponding to natural paragraphs in a text to be processed, and displaying all the generated first candidate items in the text to be processed; when the text block corresponding to the first candidate item comprises a plurality of natural paragraphs, the plurality of natural paragraphs in the text block have semantic relevance.

Specifically, the semantics corresponding to each natural paragraph in the text to be processed are obtained, and the natural paragraph with semantic relevance is determined based on the semantics corresponding to each natural paragraph, so that any natural paragraph with no semantic relevance with other natural paragraphs or a plurality of continuous natural paragraphs with semantic relevance are used as text blocks.

Further, a corresponding first candidate item is generated for the text block, so that when the user selects the first candidate item, a plurality of natural paragraphs with semantic relevance can be obtained, and the situation that the probability of the existence of the semantic relevance of continuous natural paragraphs in the conventional text is high is adapted.

In an application scenario, semantics corresponding to each natural paragraph are obtained based on a pre-trained semantic model.

In a specific application scenario, as shown in fig. 3, a first candidate in the text to be processed corresponds to two consecutive natural paragraphs with semantic relevance, and the other two first candidates respectively correspond to one natural paragraph with relatively independent semantics.

In another application mode, generating at least one first candidate item based on the semantics of text content in the text to be processed and the number of characters of the text content, and displaying all the generated first candidate items in the text to be processed; the content in the text block corresponding to the first candidate has semantic relevance, and the total characters in the text block corresponding to the first candidate are greater than a character quantity threshold.

Specifically, referring to fig. 4, fig. 4 is a schematic view of an application scenario of another embodiment corresponding to step S201 in fig. 2, the semantics of the text content in the text to be processed is obtained, and the text to be processed is divided into at least one text block with total characters greater than a threshold of the number of characters by combining the number of characters of the text content, so that the obtained text block has semantic relevance and has enough number of characters.

Further, a first candidate item (such as a gray circle in fig. 4) is set for the text block, wherein the first candidate item is not limited to being set before the natural paragraph, so that when the user selects the first candidate item, the user can obtain content with total characters larger than a threshold value of the number of characters and with semantic relevance, and more accurate abstract text can be extracted after the content in the text block is selected.

In an application scene, text segmentation position prediction is carried out based on a neural network sequence labeling frame, namely, each sentence in a text is sequentially input, whether each sentence is at a segmentation boundary is predicted, and in the training and prediction process, the segmentation boundary from the latest historical moment to the current moment of sentence segment length information characteristics, the current sentence clue word characteristics and the current sentence context semantic expression are fused, so that the effect of segmenting the text based on semantic understanding and effectively restricting paragraph length is achieved.

In another application mode, generating at least one first candidate item based on the keywords in the text to be processed and the interest group corresponding to the key sentence, and displaying all the generated first candidate items in the text to be processed; at least part of the contents in the text block corresponding to the first candidate item correspond to the same interest group.

Specifically, referring to fig. 5, fig. 5 is a schematic view of an application scenario of another embodiment corresponding to step S201 in fig. 2, in which a text to be processed is divided into words and sentences, and the text to be processed is divided into at least one text block based on a keyword and a interest group corresponding to an important sentence in the text to be processed, so that at least part of contents in the obtained text block correspond to the same interest group.

Further, a first candidate item (such as a black triangle in fig. 5) is set for the text block, wherein the first candidate item is not limited to being set before the natural paragraph, so that when the user selects the first candidate item, the user can obtain content at least partially corresponding to the same interest group, and more accurate abstract text can be extracted after the content in the text block is selected.

In an application scene, a large number of texts with key words and key sentences are collected in advance, a key word and key sentences are collected to form a combined prediction hierarchy model, the hidden sentence characteristics of context information fusion and the combined word-level key words and sentence-level key sentences are extracted, a sentence-level division prediction model fusing the key words and key sentences is formed, the trained combined prediction hierarchy model is utilized to divide the key words and key sentences, and the segmentation prediction model is utilized to divide the text to be processed into at least one text block.

In a further application mode, in response to obtaining auxiliary content matched with the text to be processed, extracting the auxiliary text from the auxiliary content, generating at least one first candidate item based on the auxiliary text, and displaying all the generated first candidate items in the text to be processed; wherein the content in the text block corresponding to the at least one first candidate is related to the auxiliary text.

Specifically, when auxiliary content which is uploaded by a user and matched with the text to be processed is obtained, auxiliary text is extracted from the auxiliary content, wherein the auxiliary content comprises at least one of pictures, handwriting and voice, so that the auxiliary text can be extracted from the pictures or the handwriting, or can be obtained after voice is converted into characters.

Further, the semantics of the auxiliary text are extracted, the text to be processed is segmented based on the semantics of the auxiliary text, at least one text block is obtained, the content of the at least one text block is matched with the semantics of the auxiliary text, a first candidate item is set for the text block, wherein the first candidate item is not limited to be set in front of a natural paragraph, so that a user can select the text block matched with the auxiliary text in the uploaded auxiliary content when selecting the first candidate item, and the matching degree of the abstract text and the user requirement is improved.

In a specific application scene, the text to be processed is text obtained after the voice related to the conference is transcribed, the auxiliary content is a picture shot in the conference scene input by the user, wherein the picture is key content required by the user, OCR recognition is performed on the picture, the text on the picture is extracted as the auxiliary text, and the text to be processed is segmented based on the semantics of the auxiliary text, so that at least one text block matched with the semantics of the auxiliary text is obtained, and is highly matched with the key content required by the user.

Further, acquiring a selected text selected by a user from the texts to be processed, including: and responding to the selected at least one first candidate item, and taking the content in the text blocks corresponding to all the selected first candidate items as the selected text.

It can be understood that when at least one first candidate item is selected, the content in the text blocks corresponding to all the selected first candidate items is taken as the selected text, and the user can select the first candidate item through any triggering form, so that the content in the text blocks corresponding to the first candidate item selected by the user is taken as the selected text.

It should be noted that any of the methods disclosed in the above implementation scenarios for generating the first candidate and providing the first candidate for the user to select, so as to obtain the content in the text block corresponding to the first candidate selected by the user, so as to obtain the selected text may also be used in the above embodiment.

S202: based on at least part of the content in the selected text, a reference text with semantic relevance to the selected text is obtained from unselected texts, wherein the unselected texts comprise texts except the selected text in the text to be processed.

Specifically, based on at least part of contents in the selected text, acquiring text which has semantic relevance with the selected text from unselected text as a reference text, wherein the unselected text is text except the selected text in the text to be processed.

In an application mode, semantic information corresponding to the selected text is obtained, content matched with the semantic information of the selected text is extracted from the unselected text, the text which has semantic relevance with the selected text in the unselected text is obtained and used as a reference text, and therefore the reference text is obtained based on the complete semantic information corresponding to the selected text, and the reliability of the obtained reference text is improved.

In another application mode, refined content is obtained from the selected text, semantic information of the refined content is obtained, content matched with the semantic information of the refined content is extracted from the unselected text, and text which has semantic relevance with the selected text in the unselected text is obtained and used as a reference text, so that the reference text is obtained based on the semantic information of the refined content, the efficiency of obtaining the reference text is improved, and the computational power consumption for obtaining the reference text is reduced.

In an application scene, the selected text and the unselected text are sent to a pre-trained semantic extraction model, the output of the semantic extraction model is obtained, semantic information corresponding to the selected text and semantic information corresponding to the unselected text are obtained, and content matched with the semantic information of the selected text is extracted from the unselected text, so that a reference text is obtained.

In another application scene, a first prompt is generated based on the selected text, the first prompt is input into a large language model, so that the large language model obtains refined content and semantic information of the refined content from the selected text based on the first prompt, unselected text is sent into a pre-trained semantic extraction model, output of the semantic extraction model is obtained, semantic information corresponding to the unselected text is obtained, and content matched with the semantic information of the refined content is extracted from the unselected text, so that a reference text is obtained.

In a specific application scenario, in order to make the user explicitly perceive that the abstract text is generated together with the reference text in addition to the selected text, the reference text is highlighted in the text to be processed after the reference text is acquired.

In another specific application scenario, in order to make the user explicitly perceive that the reference text is included in addition to the selected text and the user can select, after the reference text is acquired, a single selection box is set in front of the reference text, and all the reference texts correspond to the full selection box, so that the user can freely select the adopted reference text.

S203: and obtaining abstract text based on the selected text and the reference text.

Specifically, based on the selected text and the reference text, the key content related to the selected text is extracted to obtain the abstract text, so that the reference range for generating the abstract text is enlarged, and the probability of low accuracy of the abstract text caused by missing the content with semantic relevance to the selected text by a user is reduced.

In an application mode, extracting a first key sentence in the selected text and referring to a second key sentence in the text, and acquiring texts consisting of all the first key sentences and part of the second key sentences to obtain a abstract text, so that the abstract text is generated by taking the first key sentence as main content, and the matching degree of the abstract text and the selected text selected by a user is improved.

In another application mode, a first key sentence in the selected text is extracted, and a second key sentence in the reference text is extracted, so that a text formed by at least part of the first key sentence and at least part of the second key sentence is obtained, and a abstract text is obtained, and therefore the accuracy of the abstract text is improved by combining the first key sentence and the second key sentence.

In an application scene, setting a first weight for a selected text, setting a second weight for a reference text, extracting a first important sentence in the selected text and a second important sentence in the reference text, and obtaining a summary text based on the first weight and the first important sentence and the second weight and the second important sentence.

In a specific application scenario, the first weight of the selected text is greater than the second weight of the reference text, so that the first important sentence is assisted by the second important sentence to generate the abstract text.

In another specific application scenario, the first weight of the selected text is equal to the second weight of the reference text, so that the first important sentence and the second important sentence are integrated to generate the abstract text.

In one implementation scenario, based on at least a portion of content in the selected text, obtaining a reference text from the unselected text that has a semantic relevance to the selected text, comprising: acquiring refined content from the selected text, and acquiring a reference text which has semantic relevance with the refined content from the unselected text based on the semantic of the refined content; and generating and displaying prompt information corresponding to the reference text.

Specifically, the key content is extracted from the selected text as the refined content in the selected text, and the semantics of the refined content are obtained, so that the key content is firstly obtained from the selected text, and the semantics with higher matching degree with the selected content are obtained.

Further, based on the semantics of the refined content, extracting a reference text which has correlation with the semantics of the refined content from the unselected text, generating prompt information for the reference text, and displaying the prompt information to a user, so that the user explicitly perceives that the reference text is included in addition to the selected text to generate the abstract text.

In an application scenario, in order to facilitate the user to confirm the refined content, after obtaining the refined content from the selected text, the refined content is highlighted and a confirmation frame of the refined content is generated, and the confirmation frame can be dragged and selected, so that the user can freely select the refined content of the selected text, and further, after the user confirms the refined content on the confirmation frame, the content in the confirmation frame is used as the final refined content.

In a specific application scene, after the reference text is acquired, the position of the reference text in the text to be processed is highlighted, wherein the highlighted position can be displayed in a background color or a thickening mode of any color, so that a user perceives the position of the reference text.

Optionally, the prompt message includes at least one second candidate item corresponding to the reference text; after generating and displaying the prompt information corresponding to the reference text, the method further comprises the following steps: and in response to all the second candidates being confirmed or canceled, taking the reference text corresponding to the confirmed second candidate as a confirmation text.

Specifically, the prompt message includes at least one second candidate item corresponding to the reference text, so that the user can freely select the adopted reference text, and when all the second candidate items are confirmed or cancelled, the reference text corresponding to the confirmed second candidate item is taken as a confirmation text.

Further, obtaining the summary text based on the selected text and the reference text, including: and obtaining abstract text based on the selected text and the confirmed text.

Specifically, after the confirmation text after the formal confirmation of the user is obtained, the abstract text is obtained together based on the selected text and the confirmation text, so that the reference range of the abstract text is more matched with the requirement of the user, and the user experience is improved.

In a specific application scenario, the second candidate includes a full-choice confirmation item and a full-choice elimination item of the reference text at all positions, and a single-choice confirmation item and a single-choice elimination item of the reference text at each position, so that a user can freely select the reference text at different positions, and when the reference text at all positions is confirmed or cancelled, the reference text confirmed by the user is obtained as a confirmation text.

S204: and displaying the abstract text.

Specifically, the obtained abstract text is displayed. The display manner can be referred to the previous embodiment, and the disclosure is not repeated here.

Different from the foregoing embodiment, at least one first candidate is set in the text to be processed, after the first candidate, a text block is corresponding to the first candidate, when the user selects the first candidate, the content in the text block corresponding to the first candidate after the user selects the first candidate is used as the selected text, so that the convenience of selecting the text by the user can be improved. Therefore, the user can obtain the abstract text with high matching degree with the selected text only by selecting and selecting part of the texts perceived by the user, and the convenience of the user for obtaining the abstract text is improved.

Referring to fig. 6, fig. 6 is a flowchart of a summary generating method according to another embodiment of the present application, the method includes:

s601: in response to obtaining voice data entered by the user, the voice data is converted into text to be processed.

Specifically, when voice data input by a user is obtained, the voice data is converted into text, and the text to be processed is obtained, so that the scene that the user needs to extract the abstract from the voice data is adapted.

In an application scene, the voice data is data uploaded by a user on an application of the intelligent terminal, and the voice data uploaded by the user is converted into a text to be processed, so that the user can upload the voice data collected in any scene, and the text to be processed is obtained.

In another application scenario, the voice data are data collected during the use process of the application of the intelligent terminal, and after the application use is finished, the collected voice data are converted into the text to be processed, so that a user can obtain the text to be processed corresponding to the collected voice data during the application use process after the application use is finished.

In still another application scenario, the voice data is an intelligent terminal integrated with a text processing function and a voice acquisition function, and the voice data acquired in the use process is converted into a text to be processed when a user selects to convert the text into the intelligent terminal, so that the user can freely acquire voice in different scenarios and freely select whether to transfer or not.

In a specific application scenario, the voice data are collected during a conference process by an online conference application on the intelligent terminal, and after the online conference is finished, the collected voice data are converted into a text to be processed.

In another specific application scenario, the voice data is collected after the user triggers the recording option in the process that the intelligent terminal performs online or offline conference, and when the user selects to convert the voice into the text, the collected voice data is converted into the text to be processed.

S602: and displaying the text to be processed in the first area.

Specifically, referring to fig. 7, fig. 7 is a schematic view of an application scenario corresponding to step S602 in fig. 6, wherein a first area is a dashed box on the left side in fig. 7, and a text to be processed is displayed in the first area.

Optionally, the text to be processed includes, in addition to the text obtained by converting the speech data, a speaker corresponding to the content in the text, and when the speaker matches with the voiceprint library, the explicit speaker is prompted in the text to be processed.

S603: and acquiring at least part of contents in the text to be processed selected by the user in the first area, and displaying the selected contents in the first area as the note text in a second area, wherein the second area is not overlapped with the first area.

Specifically, referring to fig. 7, when the user selects at least part of the content in the text to be processed in the first area, the content selected by the user in the first area is displayed as a note text in the second area, where the second area is a dashed box on the right side in fig. 7, and the second area and the first area are non-overlapping, so that when the user selects the text to be processed, the content selected by the user can be displayed independently, and the content selected by the user can be displayed more intuitively.

In an application scene, performing split-screen operation on a display interface to obtain a first area and a second area which are not overlapped, displaying a text to be processed in the first area, and displaying the selected content in the first area to the second area after a user selects part of the content in the first area, so that the user can view all the content selected in the first area in the second area.

S604: and acquiring at least one note text selected by the user in the second area, and taking the selected note text in the second area as a selected text.

Specifically, when the user selects at least one note text in the second area, the selected note text in the second area is used as a selected text, so that the user can select part of content in the first area in advance as the note text to be displayed in the second area in a concentrated manner, and further screen the note text in the second area, so that the note text selected in the second area is used as a selected text confirmed by the user, the user can make comparison and reference when selecting the text, the more accurate selected text is obtained, and further the abstract text which is more matched with the user requirement can be obtained when generating the abstract text.

Optionally, the second area provides the user with an option to add or delete the note text so that the user can adjust the note text when the text to be processed is inaccurate.

S605: based at least on the selected text, a summary text is obtained.

Optionally, acquiring at least part of the content in the text to be processed selected by the user in the first area, and displaying the selected content in the first area as the note text in the second area, and further including: acquiring a time stamp matched with each note text; wherein the timestamp is associated with a speech segment to which the note text is matched in the speech data; and taking the timestamp matched with the note text corresponding to the selected text as a reference timestamp, and acquiring the reference text with semantic relevance with the selected text from the unselected text within a preset time range from the reference timestamp based on at least part of contents in the selected text.

Specifically, a timestamp of a voice segment matched with each note text in voice data is obtained, and the timestamp of each note text is saved.

Optionally, the time stamp of each note text is displayed in the second area, so that the user knows the actual corresponding generation time of each note text.

Further, when the selected text is the selected note text in the second area, the timestamp matched with the note text corresponding to the selected text is used as a reference timestamp, and the content with semantic relevance to at least part of the content in the selected text is searched from the unselected text within a preset time range from the reference timestamp, so that the reference text with semantic relevance to the selected text is obtained, the efficiency of obtaining the reference text is improved, and the processing capacity of obtaining the reference text is reduced.

In a specific application scenario, the preset time range is a transcribed text corresponding to the voice data within five minutes before and after the reference timestamp, and in other specific application scenarios, the preset time range may also be a custom time range such as ten minutes before and after, and the application is not limited in particular.

It will be appreciated that, when the reference text is obtained, the step of obtaining the summary text based at least on the selected text includes: and obtaining abstract text based on the selected text and the reference text.

In particular, the method for obtaining the abstract text based on the selected text and the reference text may refer to the above embodiment, and this will not be described in detail in the present disclosure.

S606: and displaying the abstract text.

Specifically, step S606 may refer to the related content in any of the above embodiments, which is not described herein.

Further, after presenting the abstract text, the method further comprises: generating an index relation between the abstract text and the corresponding selected text; acquiring a voice fragment of a selected text corresponding to the abstract text in voice data; based on the voice fragments, generating voice paragraphs corresponding to the abstract text.

Specifically, the selected text corresponding to the abstract text is obtained, and the index relation between the abstract text and the corresponding selected text and the index relation between the abstract text and the reference text are generated, so that a user can correspond to the source of the abstract text when selecting the corresponding abstract text, and the user can more conveniently and intuitively confirm whether the abstract text is accurate or not.

Further, based on the time stamp of the selected text corresponding to the abstract text, the corresponding voice fragment is obtained from the voice data, the voice fragment is cut and spliced, and the voice fragment corresponding to the abstract text is generated, so that a user can select to play the voice fragment corresponding to the abstract text at the abstract text, or the user can share the voice fragment corresponding to the abstract text and the source of the abstract text, so that a sharee who obtains the abstract text and the voice fragment corresponding to the abstract text can obtain more concise voice, and the inconvenience in the original voice data is protected for shared content.

Optionally, when the abstract text also corresponds to the reference text, after displaying the abstract text, the method further comprises: generating index relations between abstract texts and corresponding selected texts and reference texts; acquiring voice fragments of the selected text and the reference text corresponding to the abstract text in voice data; based on the voice fragments, generating voice paragraphs corresponding to the abstract text.

Specifically, the selected text and the reference text corresponding to the abstract text are obtained, and the index relation between the abstract text and the corresponding selected text and the reference text is generated, so that a user can correspond to the source of the abstract text when selecting the corresponding abstract text, and the user can more conveniently and intuitively confirm whether the abstract text is accurate or not.

Further, based on the time stamps of the selected text and the reference text corresponding to the abstract text, corresponding voice fragments are obtained from voice data, the voice fragments are cut and spliced, and voice fragments corresponding to the abstract text are generated, so that a user can select to play the voice fragments corresponding to the abstract text at the abstract text, or the user can share the voice fragments corresponding to the abstract text and the source of the abstract text, so that a sharee who obtains the abstract text and the voice fragments corresponding to the abstract text can obtain more concise voice, and inconvenience in the original voice data is protected for shared content.

Optionally, when the selected text corresponding to the abstract text and the reference text are matched with an explicit speaker, a prompt sound of the speaker is added to the corresponding voice segment in the finally generated voice segment, so that the user can explicitly learn the actual speaker of the corresponding voice segment.

Different from the previous embodiment, when obtaining the voice data input by the user, converting the voice data into text to obtain the text to be processed, thereby adapting to the scene that the user needs to extract the abstract from the voice data, displaying the text to be processed in the first area, displaying the content selected by the user in the first area as the note text in the second area, so that the user can pre-select part of the content as the note text in the first area to be displayed in the second area in a concentrated manner, and further screening the note text in the second area, thereby using the note text selected in the second area as the selected text confirmed by the user, enabling the user to compare and refer when selecting the text to obtain more accurate selected text, and further obtaining the abstract text which is more matched with the user needs when generating the abstract text, when the selected text is the selected note text in the second area, the timestamp matched with the note text corresponding to the selected text is used as a reference timestamp, the content with semantic relevance to at least part of the content in the selected text is searched from the unselected text within a preset time range from the reference timestamp, and the reference text with semantic relevance to the selected text is obtained, so that the efficiency of obtaining the reference text is improved, the processing capacity of obtaining the reference text is reduced, after the abstract text is obtained based on the selected text and the reference text, the abstract text is post-processed, the index relation between the abstract text and the corresponding selected text and the reference text and the voice paragraph corresponding to the abstract text are generated, and the user can conveniently review the source of the abstract text and share the abstract text and the voice paragraph corresponding to the abstract text.

Referring to fig. 8, fig. 8 is a flowchart of a summary generating method according to another embodiment of the present application, the method includes:

s801: in response to obtaining voice data entered by the user, the voice data is converted into text to be processed.

Specifically, the step is the same as the step S601, and the corresponding embodiment can be referred to, which is not repeated in the present application.

S802: and displaying the text to be processed in a third area.

Specifically, referring to fig. 9, fig. 9 is a schematic view of an application scenario corresponding to step S802 in fig. 8, in which the third area is the left area in fig. 9, and the text to be processed is shown in the third area.

S803: and acquiring a selected text selected by the user from the texts to be processed.

S804: based at least on the selected text, a summary text is obtained.

Specifically, the specific process of steps S803-S804 can refer to any one of the embodiments described above, which is not repeated herein.

In a specific application scenario, please continue to refer to fig. 9, the selected text selected by the user is represented by a solid line box, and the reference text having semantic relevance to the selected text is represented by a dashed line box, where the method is not limited to the manner described in the present scenario during actual presentation, and is only convenient for illustration and distinction. The summary text based on the selected text and the reference text is shown in fig. 9, specifically: the world is new technological revolution and industry change and storage are waiting, the international pattern is deeply evolved, and the development unbalance is not changed at all. The asia-tai is the most global economic plate with the activity and development potential and is also an important engine for worldwide economic growth. "

S805: and displaying the abstract text in a third area by utilizing the editable box.

Specifically, referring to fig. 9, the abstract text is displayed in the third area by using the editable box, so that the user can obtain the displayed abstract text at the continuous part of the text to be processed, and perform the operations of adding, deleting and modifying the displayed abstract text, thereby improving the convenience of the user in comparing the abstract text with the text to be processed.

S806: in response to detecting the user selecting the send option, content in the editable box is presented to a fourth area, wherein the fourth area is non-coincident with the third area.

Specifically, when the user selects the sending option, displaying the content in the editable box to a fourth area, wherein the fourth area is not overlapped with the third area, so that formally generated abstract text is independently displayed in an area independent of the area for displaying the text to be processed, and the user can review all finally generated abstract text together.

In a specific application scenario, please refer to fig. 10 and 11, fig. 10 is a schematic diagram of an application scenario corresponding to step S806 in fig. 8, fig. 11 is a schematic diagram of an application scenario corresponding to step S806 in fig. 8, as shown in fig. 10, when a user sends an option at a location corresponding to a dashed frame (i.e. an insert note area in fig. 10), the content in the editable box is displayed to a fourth area, as shown in fig. 11, the content in the editable box is displayed to a location in the dashed frame in fig. 11, where another text above the dashed frame in fig. 11 is a summary text that has been generated and sent to the fourth area, so that all the confirmed summary texts can be commonly displayed in the fourth area for unified review by the user.

It will be appreciated that, as shown in fig. 10, the editable box in the third area may include other options, such as copy and highlight, in addition to the send option, and the application is not limited in this regard.

Optionally, presenting the content in the editable box to the fourth area, including: acquiring a target abstract template matched with target semantics based on target semantics corresponding to the content in the editable frame; filling the content in the editable frame to the corresponding position in the target abstract template to obtain a target abstract text; and displaying the target abstract text in a fourth area.

Specifically, when the user selects to send the content in the editable box, the target semantics corresponding to the content in the editable box is obtained, the target abstract template matched with the target semantics is obtained, the content in the editable box is filled in the corresponding position in the target abstract template, so that the final target abstract text is obtained, the target abstract text is displayed in the fourth area, all the target abstract texts obtained by the user in the fourth area correspond to the respective target abstract templates, the user can distinguish the target abstract texts with different semantics, and the convenience of the user in consulting and sharing the target abstract text is improved.

In a specific application scenario, the abstract templates corresponding to different semantic types at least comprise abstract templates corresponding to discussion content, conclusion content and content to be handled, and when the target semantic corresponding to the content in the editable frame is matched with any semantic type, the content in the corresponding editable frame is added into the corresponding target abstract template.

Further, after presenting the abstract text, the method further comprises: generating an index relation between the abstract text and the corresponding selected text; acquiring a voice fragment of a selected text corresponding to the abstract text in voice data; generating a voice paragraph corresponding to the abstract text based on the voice fragment

Optionally, when the abstract text also corresponds to the reference text, after displaying the abstract text, the method further comprises: generating index relations between abstract texts and corresponding selected texts and reference texts; acquiring voice fragments of the selected text and the reference text corresponding to the abstract text in voice data; based on the voice fragments, generating voice paragraphs corresponding to the abstract text. The above process may be specifically referred to in the above embodiment, and this will not be repeated in the present application.

Different from the embodiment, the abstract text is displayed in the third area by utilizing the editable box, so that a user can obtain the displayed abstract text at the continuous part of the text to be processed, and perform adding, deleting and modifying operations on the displayed abstract text, the convenience of the user in comparing the abstract text with the text to be processed is improved, when the fact that the user selects a sending option is detected, the content in the editable box is displayed to the fourth area, wherein the fourth area is not overlapped with the third area, and therefore the formally generated abstract text is independently displayed in the area independent of the text to be processed, so that the user can review all finally generated abstract texts together, and all target abstract texts obtained in the fourth area correspond to respective target abstract templates, so that the user can distinguish the target abstract texts with different semantics, and the convenience of the user in review and sharing the target abstract text is improved.

Referring to fig. 12, fig. 12 is a schematic structural diagram of an embodiment of a summary generating device according to the present application, where the summary generating device 120 includes an obtaining module 122, a generating module 124, and a displaying module 126, where the obtaining module 122 is configured to obtain a selected text selected by a user from texts to be processed; the generating module 124 is configured to obtain abstract text based at least on the selected text; the display module 126 is used for displaying the abstract text.

Optionally, the obtaining module 122 is further configured to obtain, from the unselected text, a reference text having a semantic relevance with the selected text based on at least a part of the content in the selected text; the unselected texts comprise texts except the selected texts in the texts to be processed; the generation module is further configured to derive 124 summary text based on the selected text and the reference text.

Optionally, the presenting module 126 is further configured to present at least one first candidate item in the text to be processed; wherein each first candidate corresponds to a text block; the obtaining module 122 is further configured to, in response to at least one first candidate being selected, use contents in text blocks corresponding to all the selected first candidates as the selected text.

Optionally, the obtaining module 122 is further configured to generate at least one first candidate item based on the semantics corresponding to the natural paragraphs in the text to be processed, and display all the generated first candidate items in the text to be processed; when the text block corresponding to the first candidate item comprises a plurality of natural paragraphs, the plurality of natural paragraphs in the text block have semantic relevance.

Optionally, the obtaining module 122 is further configured to generate at least one first candidate item based on the semantics of the text content in the text to be processed and the number of characters of the text content, and display all the generated first candidate items in the text to be processed; the content in the text block corresponding to the first candidate has semantic relevance, and the total characters in the text block corresponding to the first candidate are greater than a character quantity threshold.

Optionally, the obtaining module 122 is further configured to generate at least one first candidate item based on the interest group corresponding to the keyword and the accent sentence in the text to be processed, and display all the generated first candidate items in the text to be processed; at least part of the contents in the text block corresponding to the first candidate item correspond to the same interest group.

Optionally, the obtaining module 122 is further configured to, in response to obtaining auxiliary content matching the text to be processed, extract auxiliary text from the auxiliary content, generate at least one first candidate item based on the auxiliary text, and display all the generated first candidate items in the text to be processed; wherein the content in the text block corresponding to the at least one first candidate is related to the auxiliary text.

Optionally, the obtaining module 122 is further configured to obtain refined content from the selected text, and obtain, based on the semantics of the refined content, a reference text having semantic relevance to the refined content from the unselected text; and generating and displaying prompt information corresponding to the reference text.

Optionally, the prompt message includes at least one second candidate item corresponding to the reference text; the generating module 124 is further configured to, in response to all the second candidates being confirmed or cancelled, use the reference text corresponding to the confirmed second candidate as the confirmation text; and obtaining abstract text based on the selected text and the confirmed text.

Optionally, the obtaining module 122 is further configured to, in response to obtaining voice data input by the user, convert the voice data into text to be processed; the display module 126 is further configured to display the text to be processed in the first area; the obtaining module 122 is further configured to obtain at least part of the content in the text to be processed selected by the user in the first area, and the displaying module 126 is further configured to display the selected content in the first area as the note text in the second area; wherein the second region is non-coincident with the first region; the obtaining module 122 is further configured to obtain at least one note text selected by the user in the second area, and take the selected note text in the second area as the selected text.

Optionally, the obtaining module 122 is further configured to obtain a timestamp that each note text matches; wherein the timestamp is associated with a speech segment to which the note text is matched in the speech data; the generating module 124 is further configured to use the timestamp matched with the note text corresponding to the selected text as a reference timestamp, and obtain, based on at least part of the content in the selected text, a summary text from unselected texts within a preset time range from the reference timestamp, based on the selected text and the reference text.

Optionally, the obtaining module 122 is further configured to, in response to obtaining voice data input by the user, convert the voice data into text to be processed; the display module 126 is further configured to display the text to be processed in a third area; displaying the abstract text in a third area by utilizing the editable box; responsive to detecting the user selecting the send option, displaying content in the editable box to the fourth area; wherein the fourth region is non-coincident with the third region.

Optionally, the presentation module 126 is further configured to obtain a target abstract template that matches the target semantics based on the target semantics corresponding to the content in the editable box; filling the content in the editable frame to the corresponding position in the target abstract template to obtain a target abstract text; and displaying the target abstract text in a fourth area.

Optionally, the generating module 124 is further configured to generate an index relationship between the abstract text and the corresponding selected text; acquiring a voice fragment of a selected text corresponding to the abstract text in voice data; based on the voice fragments, generating voice paragraphs corresponding to the abstract text.

Referring to fig. 13, fig. 13 is a schematic structural diagram of an embodiment of a smart terminal according to the present application, where the smart terminal 130 includes a display screen 1300, a memory 1301, and a processor 1302, the display screen 1300 and the memory 1301 are respectively coupled to the processor 1302, the display screen 1300 is at least used for displaying content to a user and for the user to select content, the memory 1301 stores program instructions (not identified), and the processor 1302 is used for executing the program instructions to implement the summary generation method described in any of the foregoing embodiments. The description of the related content is referred to the detailed description of the method embodiments, and is not repeated herein.

By the scheme, the convenience of obtaining the abstract text and the accuracy of the abstract text can be improved.

Referring to fig. 14, fig. 14 is a schematic structural diagram of an embodiment of a computer readable storage medium 140, where program instructions 1400 are stored in the computer readable storage medium 140, and when the program instructions 1400 are executed by a processor, the method for generating a digest according to any of the embodiments described above is implemented. The description of the related content is referred to the detailed description of the method embodiments, and is not repeated herein.

The units described as separate units may or may not be physically separate, and units displayed as units may or may not be physical units, may be located in one place, or may be distributed over a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiment.

In addition, each functional unit in the embodiments of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.

The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be embodied in essence or a part contributing to the prior art or all or part of the technical solution in the form of a software product stored in a storage medium, including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) or a processor (processor) to execute all or part of the steps of the methods of the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.

The foregoing description is only of embodiments of the present application, and is not intended to limit the scope of the application, and all equivalent structures or equivalent processes using the descriptions and the drawings of the present application or directly or indirectly applied to other related technical fields are included in the scope of the present application.

Claims

1. A digest generation method, the digest generation method comprising:

acquiring a selected text selected by a user from the texts to be processed;

obtaining abstract text at least based on the selected text;

and displaying the abstract text.

2. The method of claim 1, wherein before obtaining summary text based at least on the selected text, further comprising:

based on at least part of the content in the selected text, acquiring a reference text which has semantic relevance with the selected text from unselected text; wherein the unselected text comprises text other than the selected text in the text to be processed;

the obtaining abstract text at least based on the selected text comprises the following steps:

and obtaining the abstract text based on the selected text and the reference text.

3. The method according to claim 1 or 2, wherein the acquiring the selected text selected by the user before the text to be processed further comprises:

displaying at least one first candidate item in the text to be processed; wherein each first candidate corresponds to a text block;

The obtaining the selected text selected by the user from the texts to be processed comprises the following steps:

and responding to the selection of at least one first candidate item, and taking the content in the text blocks corresponding to all the selected first candidate items as the selected text.

4. A method according to claim 3, wherein said presenting at least one first candidate in said text to be processed comprises:

generating at least one first candidate item based on the semantics corresponding to the natural paragraphs in the text to be processed, and displaying all the generated first candidate items in the text to be processed;

when a plurality of natural paragraphs are included in a text block corresponding to the first candidate, semantic relativity exists among the plurality of natural paragraphs in the text block.

5. A method according to claim 3, wherein said presenting at least one first candidate in said text to be processed comprises:

generating at least one first candidate item based on the semantics of the text content in the text to be processed and the character number of the text content, and displaying all the generated first candidate items in the text to be processed;

The content in the text block corresponding to the first candidate has semantic relevance, and the total characters in the text block corresponding to the first candidate are greater than a character quantity threshold.

6. A method according to claim 3, wherein said presenting at least one first candidate in said text to be processed comprises:

generating at least one first candidate item based on the interest group corresponding to the key word and the key sentence in the text to be processed, and displaying all the generated first candidate items in the text to be processed;

at least part of the contents in the text blocks corresponding to the first candidate items correspond to the same interest group.

7. A method according to claim 3, wherein said presenting at least one first candidate in said text to be processed comprises:

extracting auxiliary text from auxiliary content in response to obtaining auxiliary content matched with the text to be processed, generating at least one first candidate item based on the auxiliary text, and displaying all the generated first candidate items in the text to be processed;

wherein content in a text block corresponding to at least one of the first candidates is related to the auxiliary text.

8. The method of claim 2, wherein the obtaining, based on at least some of the content in the selected text, a reference text from unselected text that has a semantic relevance to the selected text comprises:

acquiring refined content from the selected text, and acquiring reference text with semantic relevance to the refined content from the unselected text based on the semantics of the refined content;

and generating and displaying prompt information corresponding to the reference text.

9. The method of claim 8, wherein the hint information includes at least one second candidate corresponding to the reference text;

after the generating and displaying the prompt information corresponding to the reference text, the method further comprises the following steps:

in response to all the second candidates being confirmed or cancelled, taking the reference text corresponding to the confirmed second candidate as a confirmation text;

the step of obtaining the abstract text based on the selected text and the reference text comprises the following steps:

and obtaining the abstract text based on the selected text and the confirmed text.

10. The method of claim 1, wherein the obtaining the selected text selected by the user before the text to be processed further comprises:

Converting the voice data into the text to be processed in response to obtaining the voice data input by a user;

displaying the text to be processed in a first area;

acquiring at least part of contents in the text to be processed selected by a user in the first area, and displaying the selected contents in the first area as a note text in a second area; wherein the second region is non-coincident with the first region;

and acquiring at least one piece of the note text selected by the user in the second area, and taking the note text selected in the second area as the selected text.

11. The method of claim 10, wherein the obtaining at least part of the content in the text to be processed selected by the user in the first region, and displaying the selected content in the first region as note text in a second region, further comprises:

acquiring a time stamp matched with each note text; wherein the timestamp is related to a speech segment to which the note text is matched in the speech data;

Taking the timestamp matched with the note text corresponding to the selected text as a reference timestamp, and acquiring a reference text with semantic relevance with the selected text from the unselected text within a preset time range from the reference timestamp based on at least part of contents in the selected text;

12. The method of claim 1, wherein the obtaining the selected text selected by the user before the text to be processed further comprises:

displaying the text to be processed in a third area;

the displaying the abstract text comprises the following steps:

displaying the abstract text in the third area by utilizing an editable box;

responsive to detecting a user selection of a send option, displaying content in the editable box to a fourth area; wherein the fourth region is non-coincident with the third region.

13. The method of claim 12, wherein the exposing the content in the editable box to the fourth area comprises:

Acquiring a target abstract template matched with target semantics based on the target semantics corresponding to the content in the editable frame;

filling the content in the editable box to the corresponding position in the target abstract template to obtain a target abstract text;

and displaying the target abstract text in the fourth area.

14. The method according to claim 10 or 12, further comprising, after said presenting said summary text:

generating an index relation between the abstract text and the corresponding selected text;

acquiring a voice fragment of the selected text corresponding to the abstract text in the voice data;

and generating a voice paragraph corresponding to the abstract text based on the voice fragment.

15. A digest generation apparatus, characterized in that the digest generation apparatus comprises:

the acquisition module is used for acquiring selected texts selected by a user from the texts to be processed;

the generation module is used for obtaining abstract text at least based on the selected text;

and the display module is used for displaying the abstract text.

16. An intelligent terminal, characterized by comprising: the display screen and the memory are respectively coupled with the processor, the display screen is at least used for displaying content to a user and selecting the content for the user, the memory stores program instructions, and the processor is used for executing the program instructions to realize the abstract generating method according to any one of claims 1 to 14.

17. A computer readable storage medium having stored thereon program instructions, which when executed by a processor implement the digest generation method of any one of claims 1-14.