CN114154006A

CN114154006A - Image searching method and related device

Info

Publication number: CN114154006A
Application number: CN202111403188.0A
Authority: CN
Inventors: 杨凯航; 殷俊; 朱树磊
Original assignee: Zhejiang Dahua Technology Co Ltd
Current assignee: Zhejiang Dahua Technology Co Ltd
Priority date: 2021-11-24
Filing date: 2021-11-24
Publication date: 2022-03-08

Abstract

The application discloses an image searching method and a related device, which are used for constructing an image base library by performing dimension reduction processing on an original image in an original image library. And after the dimension reduction processing is carried out on the images to be processed, the feature quantity of each image to be processed is the same as that of the dimension reduction images in the image base. And comparing the characteristics of the image base library based on the dimension reduction characteristics to determine a dimension reduction retrieval image with higher similarity to the image to be processed, and further reducing the retrieval range. Further, a retrieval image set corresponding to the dimension reduction retrieval image is determined from the original image library, and secondary feature comparison is performed according to the image to be processed and the retrieval image set, so that the target image is determined. The above process reduces the occupation of the memory space and improves the retrieval efficiency by adopting the dimension reduction feature to carry out similarity comparison. And after the retrieval range is reduced, performing secondary feature comparison on the image features taking the image to be processed as the dimension reduction and the retrieval image set to improve the retrieval accuracy of the target image.

Description

Image searching method and related device

Technical Field

The present invention relates to the field of image processing technologies, and in particular, to an image search method and a related apparatus.

Background

With the development of science and technology, image retrieval technology is becoming more mature. The image retrieval technology is to compare the similarity of the feature vector extracted from the image to be searched with the image in the image library so as to obtain the image similar to the image to be searched.

In the related art, a linear retrieval mode is mostly adopted to retrieve image features, and the linear retrieval is to perform global feature comparison on all images in an image library through feature vectors of images to be searched. In the fields of security protection, monitoring and the like nowadays, hundreds of millions of image data are accumulated, and vectorized feature dimensions are hundreds of thousands. The method for comprehensively searching the stored images occupies a large memory space and takes a long time.

Disclosure of Invention

The embodiment of the application provides an image searching method and a related device, which are used for at least improving the efficiency of searching images.

In a first aspect, an embodiment of the present application provides an image search method, where the method includes:

performing dimensionality reduction processing on each image to be processed in the image set to be processed to obtain dimensionality reduction characteristics of the image to be searched;

screening out a first number of dimension reduction images from an image base as dimension reduction retrieval images based on the dimension reduction characteristics; the image base library is a set of dimension-reduced images subjected to dimension reduction processing on original images in an original image base, and the number of image features of each dimension-reduced image in the image base library is the same as the number of dimension-reduced features of the image to be searched;

determining a retrieval image set from the original image library based on the dimension reduction retrieval images, wherein the retrieval image set is a set of original images of each dimension reduction retrieval image;

and comparing the characteristics of the retrieval image set according to the image to be processed, and determining a target image from the retrieval image based on a comparison result.

According to the method and the device, the original image in the original image library is subjected to dimensionality reduction in advance to construct the image base library. And after the dimension reduction processing is carried out on the images to be processed, the feature quantity of each image to be processed is the same as that of the dimension reduction images in the image base. And comparing the characteristics of the image base library based on the dimension reduction characteristics to determine a dimension reduction retrieval image with higher similarity to the image to be processed, and further reducing the retrieval range. In order to improve the retrieval precision, the embodiment of the application determines a retrieval image set corresponding to the dimension reduction retrieval image from an original image library, and performs secondary feature comparison according to the image to be processed and the retrieval image set so as to determine the target image. The above process reduces the occupation of the memory space and improves the retrieval efficiency by adopting the dimension reduction feature to carry out similarity comparison. And after the retrieval range is reduced, performing secondary feature comparison on the image features taking the image to be processed as the dimension reduction and the retrieval image set to improve the retrieval accuracy of the target image.

In some possible embodiments, the screening out a first number of dimension-reduced images from the image base library as dimension-reduced retrieval images based on the dimension-reduced features includes:

dividing the image set to be processed into a plurality of groups;

for any group of images to be processed, performing feature comparison on the images to be processed and the dimension reduction images in the image base to obtain feature similarity between each image to be processed in the group of images to be processed and the dimension reduction images;

respectively selecting a second number of dimension-reduced images from each group of images to be processed based on the feature comparison result of each group of images to be processed;

and selecting a first number of dimension reduction images from the selected dimension reduction images based on the feature comparison result of the selected dimension reduction images, wherein the first number of dimension reduction images are the dimension reduction retrieval images.

When the dimension reduction features are adopted to perform feature comparison on the image base, a grouping comparison mode is adopted, the image set to be processed is divided into a plurality of groups, for each group of images to be processed, the images to be processed and the dimension reduction images in the image base are subjected to feature comparison, and a second number of dimension reduction images are screened out from each group according to comparison results. And then selecting a first number of dimension reduction images from all the screened dimension reduction images as dimension reduction retrieval images so as to improve the retrieval efficiency.

In some possible embodiments, the performing the feature comparison between the image to be processed and the dimension-reduced image in the image base includes:

dividing the dimension reduction images in the image base into a plurality of groups;

and for any one group of dimension reduction images, performing feature comparison on the image to be processed and each dimension reduction image to obtain feature similarity of the image to be processed and each dimension reduction image.

When each group of images to be processed and the dimension-reduced images in the image base are subjected to feature comparison, a grouping comparison mode can be adopted, the dimension-reduced images in the image base are divided into a plurality of groups, and for each group of dimension-reduced images, the images to be processed and each dimension-reduced image are subjected to feature comparison to obtain feature similarity of the images to be processed and each dimension-reduced image, so that the retrieval efficiency is improved.

In some possible embodiments, the selecting, based on the feature comparison result of each group of images to be processed, a second number of dimension-reduced images from each group of images to be processed respectively includes:

selecting a third number of dimension reduction images from each group of dimension reduction images based on the feature comparison result of each group of dimension reduction images;

and selecting a second number of dimension reduction images from the selected dimension reduction images based on the feature comparison result of the selected dimension reduction images.

According to the method and the device, a third number of dimension reduction images are selected from each group of dimension reduction images according to the feature comparison result of each group of dimension reduction images, a second number of dimension reduction images are selected from all the screened dimension reduction images, and the obtained second number of dimension reduction images are the images with the highest similarity to the images to be processed in all the dimension reduction images.

In some possible embodiments, each original image is provided with an image tag unique in the original image set, and each dimension-reduced image is the same as the image tag of the original image corresponding to the dimension-reduced image; determining a retrieval image set from the original image library based on the dimension-reduced retrieval image, wherein the determining comprises:

identifying an image tag of the dimension reduction retrieval image;

and acquiring an original image with the same image label as the dimension reduction retrieval image from the original image set, and determining the retrieval image set according to the original image.

According to the embodiment of the application, the original image is provided with the image label in advance, and the image label of the original image is the same as that of the dimension-reduced image after dimension reduction processing, so that after the dimension-reduced retrieval image is determined from the image base, the corresponding original image can be traced through the image label of the dimension-reduced retrieval image, and the retrieval image set is determined.

In some possible embodiments, the performing feature comparison on the retrieved image set according to the image to be processed, and determining a target image from the retrieved image based on a comparison result includes:

acquiring original image characteristics of the image to be processed, wherein the original image characteristics are image characteristics of the image to be processed without dimension reduction processing;

and comparing the characteristics of the original image with the characteristics of the retrieval image set to obtain a target image with the highest similarity with the image to be processed.

When the target image is determined, the image features of the to-be-processed image which is subjected to the dimension reduction processing are adopted to be compared with the retrieval image set, so that the loss caused by the fact that the dimension reduction features are adopted to carry out feature comparison can be avoided, and the retrieval precision is improved.

In some possible embodiments, the image base is determined according to the following:

performing dimensionality reduction processing on each original image in the original image library to obtain a dimensionality reduced image; the image feature quantity of the dimension reduction image is smaller than that of the original image, and the ratio of the image feature quantity of the dimension reduction image to that of the original image is a preset ratio;

and carrying out quantization processing on the image characteristics of the dimension reduction image so as to enable the byte number of the image characteristics of each dimension reduction image in the image base to be smaller than the byte number before quantization processing.

According to the method and the device, the original image library is subjected to dimension reduction processing to obtain the dimension reduction image, and the image characteristics of the dimension reduction image are subjected to quantization processing to reduce the byte number of the image characteristics of the dimension reduction image, so that the data size of the image base composed of the dimension reduction image is greatly reduced compared with the data size of the original image library, and the occupation of the memory space is further reduced.

In some possible embodiments, before the feature comparing the image to be processed with the dimension-reduced image in the image base, the method further includes:

and carrying out inverse quantization processing on the dimension reduction image, wherein the image characteristics of the dimension reduction image after inverse quantization processing are the same as the characteristic types of the image characteristics of the original image corresponding to the dimension reduction image.

The dimension reduction image in the embodiment of the application has smaller memory occupation after quantization processing, and the dimension reduction image needs to be subjected to inverse quantization before the feature comparison between the image to be processed and the dimension reduction image is performed, so that the retrieval precision is ensured.

In some possible embodiments, the image feature type of the dimension-reduced image after the quantization processing is performed on the image features of the dimension-reduced image is an integer type.

After the image characteristics of the dimension-reduced image are subjected to quantization processing, the image characteristic type is changed into an integer type int, so that the number of bytes of the image characteristics is 2, and the memory occupation is further reduced.

In a second aspect, an embodiment of the present application provides an image search apparatus, including:

the dimensionality reduction feature acquisition module is used for performing dimensionality reduction on each image to be processed in the image set to be processed to acquire dimensionality reduction features of the image to be searched;

the dimension reduction image screening module is used for screening a first number of dimension reduction images from the image base as dimension reduction retrieval images based on the dimension reduction characteristics; the image base library is a set of dimension-reduced images subjected to dimension reduction processing on original images in an original image base, and the number of image features of each dimension-reduced image in the image base library is the same as the number of dimension-reduced features of the image to be searched;

a retrieval image determining module, which determines a retrieval image set from the original image library based on the dimension reduction retrieval images, wherein the retrieval image set is a set of original images of each dimension reduction retrieval image;

and the target image determining module is used for comparing the characteristics of the retrieval image set according to the image to be processed and determining a target image from the retrieval image based on a comparison result.

In some possible embodiments, the screening out, based on the dimension reduction features, a first number of dimension reduction images from an image base as dimension reduction retrieval images is performed, and the dimension reduction image screening module is configured to:

dividing the image set to be processed into a plurality of groups;

In some possible embodiments, performing the feature comparison between the image to be processed and the dimension-reduced image in the image base library, where the dimension-reduced image screening module is configured to:

In some possible embodiments, the performing selects a second number of dimension-reduced images from each group of images to be processed based on the feature comparison result of each group of images to be processed, and the dimension-reduced image screening module is configured to:

In some possible embodiments, each original image is provided with an image tag unique in the original image set, and each dimension-reduced image is the same as the image tag of the original image corresponding to the dimension-reduced image; performing the determining of the set of search images from the original image library based on the dimension-reduced search image, the search image determination module being configured to:

identifying an image tag of the dimension reduction retrieval image;

In some possible embodiments, the feature comparison is performed on the retrieval image set according to the image to be processed, and a target image is determined from the retrieval images based on a comparison result, and the target image determination module is configured to:

In some possible embodiments, before performing the feature comparison between the image to be processed and the dimension-reduced image in the image base, the dimension-reduced image screening module is further configured to:

In a third aspect, an embodiment of the present application further provides an electronic device, including at least one processor; and a memory communicatively coupled to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute a method for searching a profile of a target object provided by the embodiment of the application.

In a fourth aspect, an embodiment of the present application further provides a computer storage medium, where a computer program is stored, where the computer program is used to enable a computer to execute an archive search method for a target object provided in the embodiment of the present application.

Additional features and advantages of the application will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the application. The objectives and other advantages of the application may be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments of the present application will be briefly described below, and it is obvious that the drawings described below are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

Fig. 1 is a schematic view of an application scenario shown in an embodiment of the present application;

FIG. 2a is a flowchart illustrating an overall image searching method according to an embodiment of the present disclosure;

fig. 2b is a schematic diagram illustrating reduction of memory space occupation according to an embodiment of the present application;

FIG. 2c is a schematic diagram illustrating determining a dimension-reduced search image according to an embodiment of the present application;

FIG. 2d is a schematic diagram showing a comparison of features in an embodiment of the present application;

fig. 3 is a block diagram of an image search apparatus 300 according to an embodiment of the present application;

fig. 4 is a schematic diagram of an electronic device according to an embodiment of the present application.

Detailed Description

The technical solutions in the embodiments of the present application will be described in detail and clearly with reference to the accompanying drawings. In the description of the embodiments of the present application, unless otherwise specified, "a face will mean or means, for example, a/B may mean a or B; "and/or" in the text is only an association relationship describing an associated object, and means that three relationships may exist, for example, a and/or B may mean: three cases of a alone, a and B both, and B alone exist, and in addition, "a plurality" means two or more than two in the description of the embodiments of the present application.

In the description of the embodiments of the present application, the term "plurality" means two or more unless otherwise specified, and other terms and the like should be understood similarly, and the preferred embodiments described herein are only for the purpose of illustrating and explaining the present application, and are not intended to limit the present application, and features in the embodiments and examples of the present application may be combined with each other without conflict.

To further illustrate the technical solutions provided by the embodiments of the present application, the following detailed description is made with reference to the accompanying drawings and the detailed description. Although the embodiments of the present application provide method steps as shown in the following embodiments or figures, more or fewer steps may be included in the method based on conventional or non-inventive efforts. In steps where no necessary causal relationship exists logically, the order of execution of the steps is not limited to that provided by the embodiments of the present application. The method can be executed in the order of the embodiments or the method shown in the drawings or in parallel in the actual process or the control device.

The image retrieval technology is widely applied, for example, in case of solving a case in the public security industry, a large number of human images in an image library need to be retrieved according to suspect images of which the identities are not confirmed so as to obtain identity information of the suspect. In the related art, after the image features of the image to be searched are extracted, the image features are pulled into a memory in a full amount. And comparing the features with the stored images (namely the image library) and outputting the images with the similarity higher than a preset threshold. However, the method of searching the stored image comprehensively occupies a large amount of memory space and takes a long time.

In order to solve the above problems, the inventive concept of the present application is: according to the method and the device, the original image in the original image library is subjected to dimensionality reduction in advance to construct the image base library. And after the dimension reduction processing is carried out on the images to be processed, the feature quantity of each image to be processed is the same as that of the dimension reduction images in the image base. Therefore, the feature comparison can be carried out on the image base on the basis of the dimension reduction features so as to determine the dimension reduction retrieval image with higher similarity to the image to be processed. Since the feature comparison by using the dimension reduction features has a certain loss compared with the feature of the image without dimension reduction, in order to improve the retrieval accuracy, the embodiment of the application determines the retrieval image set corresponding to the dimension reduction retrieval image from the original image library, and performs secondary feature comparison according to the image to be processed and the retrieval image set, so as to determine the target image. The above process reduces the occupation of the memory space by adopting the dimension reduction characteristic to carry out similarity comparison, thereby improving the retrieval efficiency.

First, an image to be processed and an original image in the embodiment of the present application are explained:

the image to be processed in the embodiment of the present application may be any image with a retrieval requirement, such as but not limited to images including a human face image, a human body image, an animal image, a plant image, a building image, a vehicle image, a commodity or an article image, and the like; the original image is a Video frame image captured by using, for example, a Network Video Recorder (NVR), or a single frame image captured by using other image capturing devices, such as but not limited to a face image, a human body image, an animal image, a plant image, a building image, a vehicle image, a commodity or an article image, and the like; in the following of the embodiments of the present application, a face image is taken as a specific example of an image to be processed, and the image searching method and the related apparatus provided by the present application are further described.

The archive search method for the target object in the embodiment of the present application is described in detail below with reference to the drawings.

Referring to fig. 1, a schematic diagram of an application environment according to an embodiment of the present application is shown.

As shown in fig. 1, the application environment may include, for example, a network 10, a server 20, at least one terminal device 30, and a database 40. Among them, the terminal device 30 may include a smartphone 30_1, a desktop computer 30_2, and a notebook computer 30_ n shown in fig. 1.

In the application scenario shown in fig. 1, the terminal device 30 transmits an image to be searched to the server 20 through the network 10 by inputting a plurality of images to be searched into the terminal device 30 by the user. The server 20 determines the dimension reduction features of the image to be processed by performing dimension reduction processing on the image to be processed, so that the dimension reduction features are the same as the number of features of each dimension reduction image in the image base pre-stored in the database 40.

In some possible embodiments, the server 20 searches the image base pre-stored in the database 40 based on the dimension reduction feature, and retrieves the first n images with the highest similarity, where n is an integer. Further, determining original images of the n images, wherein each original image is an image before dimension reduction processing of the n images.

In some possible embodiments, the image features of the image to be processed without dimension reduction processing are adopted to perform feature comparison with each original image so as to determine the target image with the highest similarity.

After introducing an application scenario to which the technical solution of the present application is applicable, a detailed description is provided below with reference to the accompanying drawings for an image search method provided in the embodiment of the present application, specifically as shown in fig. 2a, including the following steps:

step 201: performing dimensionality reduction processing on each image to be processed in the image set to be processed to obtain dimensionality reduction characteristics of the image to be searched;

the following description will be given by taking a 512-dimensional, float-point type standard image as the image to be processed and the original image in the original image library. It should be understood that the above definitions of the number and types of the image features are only used for convenience of describing the technical solutions provided in the present application, and do not limit the application scope of the present application.

The dimension reduction processing is performed on the image to be processed in advance, and the dimension of the image to be processed is reduced to 256 dimensions, so that the memory occupied by the image to be processed is reduced by one time.

Step 202: screening out a first number of dimension reduction images from an image base as dimension reduction retrieval images based on the dimension reduction characteristics; the image base library is a set of dimension-reduced images subjected to dimension reduction processing on original images in an original image base, and the number of image features of each dimension-reduced image in the image base library is the same as the number of dimension-reduced features of the image to be searched;

the image retrieval technology needs to search an image library for a target image similar to an image to be searched according to the image to be searched. It has been mentioned above that, for the convenience of description, a face image is taken as an image to be searched in the embodiment of the present application. Correspondingly, the original image library in the embodiment of the application is a video frame image library or a portrait library constructed based on the image acquisition device in the fields of security and the like.

The method and the device for searching the image feature of the original image library perform dimension reduction processing on each original image in the original image library in advance to obtain the dimension reduction images, wherein the image feature quantity of each dimension reduction image is the same as the dimension reduction feature quantity of the image to be searched, the image feature quantity of the dimension reduction images is smaller than the image feature quantity of the original images, and the ratio of the image feature quantity of the dimension reduction images to the image feature quantity of the original images is a preset ratio. Specifically, as shown in fig. 2b, for example, if the original image is 512 dimensions, the dimension-reduced image is 256 dimensions. Further, the dimension-reduced image is subjected to quantization processing, the purpose of the quantization processing is to change the image feature type of the dimension-reduced image, specifically, if the original image is of float type, the dimension-reduced image obtained after the dimension-reduction processing is also of float type. The dimension-reduced image is changed into an int type by performing quantization processing on the dimension-reduced image of float type, and thus, the dimension-reduced image is reduced from 8 bytes to 2 bytes.

Through the processing, when the original 512-dimensional and float-type original images are changed into 256-dimensional and int-type dimension-reduced images, the characteristic dimension of the images is reduced to one half of the original dimension, and the memory occupation is reduced by one half. Furthermore, the image characteristics of the original float-type dimension-reduced image are changed into int-type image, the byte is reduced to one fourth of the original image, and the memory occupation is reduced by 4 times. Therefore, the image base library constructed according to the dimension reduction images is reduced by 8 times compared with the memory occupied by the original image library. In consideration of the fact that most face feature retrieval in practical application is a batch processing flow. That is, a plurality of facial images captured by NVR (network video recorder) or other electronic devices are stored in the memory, and the image library is retrieved in a batch comparison manner to determine a target image corresponding to each image to be processed (i.e., an image in the image library with the highest similarity to each image to be processed).

Based on the method and the device, after the dimension reduction characteristics of the to-be-processed images in the to-be-processed image set are obtained, the to-be-processed images can be quantized into an int type and stored in the memory, so that the occupied space of the memory is reduced. When the feature comparison is carried out, firstly, the image set to be processed is divided into a plurality of groups, and for any group of images to be processed, the images to be processed and the dimension reduction images in the image base are subjected to feature comparison to obtain the feature similarity of each image to be processed in the group of images to be processed and the dimension reduction images. And then, respectively selecting a second number of dimension reduction images from each group of images to be processed based on the feature comparison result of each group of images to be processed. And selecting a first number of dimension reduction images from the selected dimension reduction images as dimension reduction retrieval images based on the feature comparison result of the selected dimension reduction images.

Specifically, as shown in fig. 2c, for example, 100 images in the image set to be processed are divided into 10 groups, 10 groups in total, and the numbers are 1 to 10. And then, carrying out feature comparison on each group of images to be processed, and selecting the first 10 dimensionality reduction images with the highest similarity from the images according to a feature comparison result. And selecting 10 images with the highest similarity from the 10 selected dimension-reduced images (total 100 images) in each group as dimension-reduced retrieval images.

In addition, considering that the search rate can be effectively improved by a grouping comparison mode, when the feature comparison is performed on the dimension reduction images in the image base of the image to be processed, the dimension reduction images in the image base can be divided into a plurality of groups, and for any one group of dimension reduction images, the feature comparison is performed on the image to be processed and each dimension reduction image, so that the feature similarity between the image to be processed and each dimension reduction image is obtained. And then, selecting a third number of dimension reduction images from each group of dimension reduction images based on the feature comparison result of each group of dimension reduction images. And selecting a second number of dimension reduction images from the selected dimension reduction images based on the feature comparison result of the selected dimension reduction images.

Specifically, as shown in fig. 2d, for example, a total of 100 images in the image set to be processed, and a total of 10000 dimensionality reduction images in the image base library are obtained. Firstly, dividing 100 images to be processed into 10 groups, wherein the total number of the 10 groups is 1-10. Then, dividing the dimension reduction images into 1000 groups, wherein the 10 groups are counted, and the labels are a-j. Taking the group 10 of the images to be processed as an example, when performing feature comparison on each group of the images to be processed, respectively performing feature comparison with each group a to j, screening the first 10 images with the highest similarity according to the feature comparison result of each group, screening 100 images in total, and selecting 10 images with the highest similarity from the 100 screened images as the 10 dimension-reduced images with the highest similarity obtained by feature comparison of the group of the images to be processed.

It should be noted that, the dimension reduction features of the image base and the to-be-processed image stored in the memory are both int type, and in order to improve the retrieval accuracy, inverse quantization processing should be performed on the to-be-processed image and the dimension reduction image before feature comparison is performed, that is, int type is changed into float type. Because the embodiment of the application performs the feature comparison in groups, the calculation amount (namely, the comparison of each group of features) of each time is greatly reduced compared with the global feature comparison, the resource occupation ratio of inverse quantization processing is very small, and the calculation efficiency can be effectively improved compared with the global feature comparison through testing.

Step 203: determining a retrieval image set from the original image library based on the dimension reduction retrieval images, wherein the retrieval image set is a set of original images of each dimension reduction retrieval image;

in the embodiment of the application, each original image is provided with a unique image label in the original image set, and each dimension reduction image is the same as the image label of the original image corresponding to the dimension reduction image. That is, the image label of the dimension-reduced image after the dimension reduction processing of the original image is the same as that of the original image. Since the aforementioned dimension reduction retrieval image is obtained by screening from the image base, the essence of the image is still dimension reduction image. Based on this, after the dimension reduction retrieval image is obtained, the original image corresponding to each dimension reduction retrieval image can be determined by inquiring the image label of the dimension reduction retrieval image and retrieving the original image library based on the image label, and each original image obtained by retrieval is a retrieval image set.

Step 204: and comparing the characteristics of the retrieval image set according to the image to be processed, and determining a target image from the retrieval image based on a comparison result.

Considering that certain loss exists when the dimension reduction features are adopted for feature comparison, the retrieval precision is lower than that of the image features without dimension reduction. Therefore, when the retrieval image set is subjected to feature comparison according to the image to be processed and the target image is determined based on the comparison result, the original image feature of the image to be processed is determined at first, and the original image feature is the image feature of the image to be processed which is not subjected to dimension reduction processing. And then, comparing the characteristics of the original image with the characteristics of the retrieval image set to obtain a target image with the highest similarity with the image to be processed. In the retrieval stage, the retrieval time consumption is negligible due to the great reduction of the compared data volume, and the retrieval time consumption is mainly reflected in the database query stage, namely the process of acquiring the retrieval image set. By employing parallel and some database query operation optimization, this phase time consumption can be effectively reduced.

Therefore, the process of the embodiment of the application firstly adopts the dimension reduction feature to carry out similarity comparison so as to reduce the occupation of the memory space and improve the retrieval efficiency. And then, after the retrieval range is reduced (namely after the dimension reduction retrieval image is determined from the original image library), performing secondary feature comparison on the image features taking the image to be processed as the dimension reduction and the retrieval image set to improve the retrieval accuracy of the target image.

Based on the same inventive concept, an embodiment of the present application provides a face feature search apparatus 300, specifically as shown in fig. 3, including: a dimension reduction feature obtaining module 301, configured to perform dimension reduction processing on each image to be processed in the image set to be processed, and obtain a dimension reduction feature of the image to be searched;

a dimension reduction feature obtaining module 301, configured to perform dimension reduction processing on each to-be-processed image in the to-be-processed image set, so as to obtain a dimension reduction feature of the to-be-searched image;

the dimension reduction image screening module 302 is used for screening a first number of dimension reduction images from the image base as dimension reduction retrieval images based on the dimension reduction characteristics; the image base library is a set of dimension-reduced images subjected to dimension reduction processing on original images in an original image base, and the number of image features of each dimension-reduced image in the image base library is the same as the number of dimension-reduced features of the image to be searched;

a retrieval image determining module 303, configured to determine a retrieval image set from the original image library based on the dimension-reduced retrieval images, where the retrieval image set is a set of original images of each dimension-reduced retrieval image;

and the target image determining module 304 is configured to perform feature comparison on the retrieved image set according to the image to be processed, and determine a target image from the retrieved image based on a comparison result.

In some possible embodiments, performing the screening out the first number of dimension-reduced images from the image base library based on the dimension-reduced features is configured to:

dividing the image set to be processed into a plurality of groups;

In some possible embodiments, performing the feature comparison between the image to be processed and the dimension-reduced image in the image base, the dimension-reduced image screening module 302 is configured to:

In some possible embodiments, the performing selects a second number of dimension-reduced images from each group of images to be processed based on the feature comparison result of each group of images to be processed, and the dimension-reduced image filtering module 302 is configured to:

In some possible embodiments, each original image is provided with an image tag unique in the original image set, and each dimension-reduced image is the same as the image tag of the original image corresponding to the dimension-reduced image; performing the determining of the set of retrieved images from the original image library based on the dimension-reduced retrieved images, the retrieved images determining module 303 is configured to:

identifying an image tag of the dimension reduction retrieval image;

In some possible embodiments, the feature comparison is performed on the retrieved image set according to the image to be processed, and a target image is determined from the retrieved image based on a comparison result, and the target image determination module 304 is configured to:

In some possible embodiments, before performing the feature comparison between the image to be processed and the dimension-reduced image in the image base, the dimension-reduced image screening module 302 is further configured to:

The electronic device 130 according to this embodiment of the present application is described below with reference to fig. 4. The electronic device 130 shown in fig. 4 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present application.

As shown in fig. 4, the electronic device 130 is represented in the form of a general electronic device. The components of the electronic device 130 may include, but are not limited to: the at least one processor 131, the at least one memory 132, and a bus 133 that connects the various system components (including the memory 132 and the processor 131).

Bus 133 represents one or more of any of several types of bus structures, including a memory bus or memory controller, a peripheral bus, a processor, or a local bus using any of a variety of bus architectures.

The memory 132 may include readable media in the form of volatile memory, such as Random Access Memory (RAM)1321 and/or cache memory 1322, and may further include Read Only Memory (ROM) 1323.

Memory 132 may also include a program/utility 1325 having a set (at least one) of program modules 1324, such program modules 1324 including, but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may comprise an implementation of a network environment.

The electronic device 130 may also communicate with one or more external devices 134 (e.g., keyboard, pointing device, etc.), with one or more devices that enable a user to interact with the electronic device 130, and/or with any devices (e.g., router, modem, etc.) that enable the electronic device 130 to communicate with one or more other electronic devices. Such communication may occur via input/output (I/O) interfaces 135. Also, the electronic device 130 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network, such as the internet) via the network adapter 136. As shown, network adapter 136 communicates with other modules for electronic device 130 over bus 133. It should be understood that although not shown in the figures, other hardware and/or software modules may be used in conjunction with electronic device 130, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.

In some possible embodiments, aspects of an image search method provided by the present application may also be implemented in the form of a program product including program code for causing a computer device to perform the steps of an image search method according to various exemplary embodiments of the present application described above in this specification when the program product is run on the computer device.

The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

The program product for image search of the embodiments of the present application may employ a portable compact disc read only memory (CD-ROM) and include program codes, and may be executed on an electronic device. However, the program product of the present application is not limited thereto, and in this document, a readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.

A readable signal medium may include a propagated data signal with readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A readable signal medium may also be any readable medium that is not a readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

Program code embodied on a readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

Program code for carrying out operations of the present application may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the consumer electronic device, partly on the consumer electronic device, as a stand-alone software package, partly on the consumer electronic device and partly on a remote electronic device, or entirely on the remote electronic device or server. In the case of remote electronic devices, the remote electronic devices may be connected to the consumer electronic device through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external electronic device (e.g., through the internet using an internet service provider).

It should be noted that although several units or sub-units of the apparatus are mentioned in the above detailed description, such division is merely exemplary and not mandatory. Indeed, the features and functions of two or more units described above may be embodied in one unit, according to embodiments of the application. Conversely, the features and functions of one unit described above may be further divided into embodiments by a plurality of units.

Further, while the operations of the methods of the present application are depicted in the drawings in a particular order, this does not require or imply that these operations must be performed in this particular order, or that all of the illustrated operations must be performed, to achieve desirable results. Additionally or alternatively, certain steps may be omitted, multiple steps combined into one step execution, and/or one step broken down into multiple step executions.

As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and block diagrams, and combinations of flows and blocks in the flow diagrams and block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and block diagram block or blocks.

While the preferred embodiments of the present application have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all alterations and modifications as fall within the scope of the application.

It will be apparent to those skilled in the art that various changes and modifications may be made in the present application without departing from the spirit and scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include such modifications and variations as well.

Claims

1. An image search method, characterized in that the method comprises:

2. The method of claim 1, wherein the screening out a first number of dimension-reduced images from an image base as dimension-reduced retrieval images based on the dimension-reduced features comprises:

dividing the image set to be processed into a plurality of groups;

3. The method according to claim 2, wherein the performing the feature comparison between the image to be processed and the dimension-reduced image in the image base includes:

4. The method according to claim 3, wherein the selecting a second number of dimension-reduced images from each group of images to be processed based on the feature comparison result of each group of images to be processed comprises:

5. The method according to claim 1, wherein each original image is provided with an image label unique in an original image set, and each dimension-reduced image is identical to the image label of the original image corresponding to the dimension-reduced image; determining a retrieval image set from the original image library based on the dimension-reduced retrieval image, wherein the determining comprises:

identifying an image tag of the dimension reduction retrieval image;

6. The method of claim 1, wherein the performing feature comparison on the retrieved image set according to the image to be processed and determining a target image from the retrieved image based on the comparison result comprises:

7. The method of any of claims 1-6, wherein the image base is determined according to:

8. The method according to claim 7, wherein before the feature comparison between the image to be processed and the dimension-reduced image in the image base, the method further comprises:

9. The method according to claim 7, wherein the image feature type of the dimension-reduced image after the quantization processing is performed on the image feature of the dimension-reduced image is an integer type.

10. An image search apparatus, characterized in that the apparatus comprises:

11. An electronic device comprising at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-9.

12. A computer storage medium, characterized in that the computer storage medium stores a computer program for causing a computer to perform the method according to any one of claims 1-9.