[go: up one dir, main page]
More Web Proxy on the site http://driver.im/

CN116759061B - Physical examination project recommendation system based on personal demands - Google Patents

Physical examination project recommendation system based on personal demands Download PDF

Info

Publication number
CN116759061B
CN116759061B CN202311035553.6A CN202311035553A CN116759061B CN 116759061 B CN116759061 B CN 116759061B CN 202311035553 A CN202311035553 A CN 202311035553A CN 116759061 B CN116759061 B CN 116759061B
Authority
CN
China
Prior art keywords
module
physical examination
input end
output end
wavelet
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202311035553.6A
Other languages
Chinese (zh)
Other versions
CN116759061A (en
Inventor
徐洪霞
朱群仙
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
JIANYANG CITY PEOPLE'S HOSPITAL
Original Assignee
JIANYANG CITY PEOPLE'S HOSPITAL
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by JIANYANG CITY PEOPLE'S HOSPITAL filed Critical JIANYANG CITY PEOPLE'S HOSPITAL
Priority to CN202311035553.6A priority Critical patent/CN116759061B/en
Publication of CN116759061A publication Critical patent/CN116759061A/en
Application granted granted Critical
Publication of CN116759061B publication Critical patent/CN116759061B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
    • G16H40/20—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/25—Fusion techniques
    • G06F18/253—Fusion techniques of extracted features
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/06—Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice
    • G10L15/063—Training
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/08—Speech classification or search
    • G10L15/16—Speech classification or search using artificial neural networks
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/20—Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/26—Speech to text systems
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
    • G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
    • G10L21/0208—Noise filtering
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
    • G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Computational Linguistics (AREA)
  • Evolutionary Computation (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Human Computer Interaction (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Biomedical Technology (AREA)
  • General Health & Medical Sciences (AREA)
  • Evolutionary Biology (AREA)
  • Mathematical Physics (AREA)
  • Biophysics (AREA)
  • Computing Systems (AREA)
  • Software Systems (AREA)
  • Molecular Biology (AREA)
  • Business, Economics & Management (AREA)
  • General Business, Economics & Management (AREA)
  • Signal Processing (AREA)
  • Medical Informatics (AREA)
  • Public Health (AREA)
  • Epidemiology (AREA)
  • Primary Health Care (AREA)
  • Quality & Reliability (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention provides a physical examination project recommending system based on personal requirements, which belongs to the technical field of man-machine interaction systems.

Description

Physical examination project recommendation system based on personal demands
Technical Field
The invention relates to the technical field of human-computer interaction systems, in particular to a physical examination project recommendation system based on personal requirements.
Background
With the promotion of public health consciousness, the number of people attending physical examination is gradually increased, doctors and nurses are limited to hospitals, and due to different human body examination demands, the required physical examination items are different, if the required physical examination items are acquired through queuing or registration, the physical examination time is increased for the physical examination personnel, and the working intensity is increased for medical workers.
The existing physical examination guide and examination system is used for guiding physical examination staff to execute according to the procedure by pushing the physical examination procedure, realizing the standard procedure and reducing the physical examination time of the physical examination staff, but because the requirements of each person are different, the physical examination staff still need to guide by the medical care staff, thereby determining physical examination items and not reducing the workload of medical care work.
Disclosure of Invention
Aiming at the defects in the prior art, the physical examination item recommending system based on personal requirements solves the problem that the existing system for automatically acquiring physical examination items required by physical examination personnel is lacking.
In order to achieve the aim of the invention, the invention adopts the following technical scheme: a physical examination item recommendation system based on personal needs, comprising: the device comprises a voice acquisition denoising unit, a voice recognition unit, a keyword extraction unit, a keyword matching unit and a physical examination item pushing unit;
the voice acquisition denoising unit is used for acquiring voice signals of physical examination personnel and performing wavelet transformation denoising on the voice signals to obtain denoised voice signals; the voice recognition unit is used for recognizing the denoising voice signal to obtain text information; the keyword extraction unit is used for extracting physical examination keywords from the text information; the keyword matching unit is used for matching the physical examination keywords with the descriptions of each physical examination item and selecting the physical examination item successfully matched; the physical examination item pushing unit is used for pushing the physical examination items successfully matched with the corresponding physical examination processes.
Further, the voice acquisition denoising unit includes: a wavelet decomposition subunit, a wavelet coefficient selection subunit, and a reconstruction subunit;
the wavelet decomposition subunit is used for performing wavelet decomposition on the voice signal to obtain wavelet coefficients;
the wavelet coefficient selection subunit is used for updating the wavelet coefficient according to the threshold value to obtain an updated wavelet coefficient;
the reconstruction subunit is used for carrying out reconstruction processing on the updated wavelet coefficient to obtain a denoising voice signal.
Further, the expression of the updated wavelet coefficients is:
wherein ,for the updated->Layer wavelet coefficients, < >>For the%>Layer wavelet coefficients, < >>To update the coefficients.
The beneficial effects of the above further scheme are: the functions of the existing updated wavelet coefficients are segment functions, and the functions are discontinuous at the joint points of the segment functions, so that the precision of the updated wavelet coefficients is not high, therefore, the invention utilizes the continuous smoothing function with the function value between-1 and 1A new expression for updating the wavelet coefficients is constructed, so that the whole definition domain range is smooth, the wavelet coefficients can be precisely removed, and the denoising precision is improved.
Further, the update coefficientThe expression of (2) is:
wherein ,is->Threshold value of layer wavelet coefficient, < >>For the%>Layer wavelet coefficients.
The beneficial effects of the above further scheme are: updating coefficientsFollowing->Layer wavelet coefficients->And threshold->The adaptive change ensures that the function of the wavelet coefficient has good transition near the threshold value.
Further, the firstThreshold value of layer wavelet coefficient->The expression of (2) is:
wherein ,is a proportional coefficient->Is->The>Wavelet coefficient value,/">Is the number of wavelet coefficient values, +.>For the length of the speech signal, < >>For wavelet decomposition scale, +.>As a logarithmic function.
The beneficial effects of the above further scheme are: the invention utilizes the firstThe average value of wavelet coefficient values in the wavelet coefficients of the layers is used for estimating the threshold value, and the proportional coefficient is set for adjusting the threshold value.
Further, the voice recognition unit includes: the system comprises a convolution module, a residual error module, a first LSTM module, a second LSTM module, an attention module, a CNN network and a CTC classifier;
the input end of the convolution module is connected with the input end of the first LSTM module and is used as the input end of the voice recognition unit; the output end of the convolution module is connected with the input end of the residual error module; the input end of the second LSTM module is connected with the output end of the first LSTM module; the input end of the attention module is respectively connected with the output end of the residual error module and the output end of the second LSTM module, and the output end of the attention module is connected with the input end of the CNN network; the input end of the CTC classifier is connected with the output end of the CNN network, and the output end of the CTC classifier is used as the output end of the voice recognition unit.
The beneficial effects of the above further scheme are: the invention utilizes two paths to extract the characteristics of the denoising voice signal respectively, enriches the characteristic quantity, utilizes the LSTM module to have time memory, can better consider the history characteristics, the residual error module fuses the deep level and the shallow level sub-characteristics, the output of the LSTM module and the output of the residual error module are weighted and fused in the attention module, the processing is carried out according to the significance degree of each characteristic, the weight is applied to each characteristic in a self-adaption manner, the weighted and fused characteristics are input into the CNN network to carry out the deep extraction of the characteristics, and then the text information is output through the CTC classifier.
Further, the residual module includes: the system comprises a first convolution sub-module, a second convolution sub-module, a third convolution sub-module, an adder and a multiplier;
the input end of the first convolution sub-module is respectively connected with the input end of the third convolution sub-module and the first input end of the multiplier, and the output end of the first convolution sub-module is connected with the input end of the second convolution sub-module; the first input end of the adder is connected with the output end of the second convolution sub-module, the second input end of the adder is connected with the output end of the third convolution sub-module, and the output end of the adder is connected with the second input end of the multiplier; the output end of the multiplier is used as the output end of the residual error module.
The beneficial effects of the above further scheme are: according to the invention, the features processed by the two convolution sub-modules and the features processed by one convolution sub-module are added through the adder, so that the information quantity is improved, and then the multiplier is used for multiplying the input features, so that the problem of gradient disappearance can be solved, and the shallow features can be fused.
Further, the expression of the attention module is:
wherein ,for the output of the attention module, +.>For Concat splice operation, < >>Is->Weight of->For the output of the residual block,/>Is->Weight of->Is the output of the second LSTM module.
Further, the saidWeight of +.>The expression of (2) is:
wherein ,as an exponential function based on natural constants, < +.>To activate the function +.>For global pooling processing,/->Is the output of the residual error module;
the saidWeight of +.>The expression of (2) is:
wherein ,is the output of the second LSTM module.
The beneficial effects of the above further scheme are: in the invention respectively to and />Different weights are given, and splicing is carried out, so that feature fusion is realized, and the method is characterized in that +_> and />According to-> and />The self situation is calculated, and the self-adaption attention degree of the remarkable characteristics is improved.
Further, the loss function of the speech recognition unit is:
wherein ,for loss function->Is->Status function at sub-training>The true class in the denoised speech signal samples is equal to the class +.>When 1 is taken, if not, 0 is taken, +.>As an exponential function based on natural constants, < +.>Is->Predictive probability of speech recognition unit during secondary training, < >>For the current training times, +.>For a small scale adjacent to the training times, +.>For the total number of adjacent exercises>Is the number of categories.
The beneficial effects of the above further scheme are: in the present inventionWhen equal to 1>The closer to 0, the more the prediction differs from the tag, so the present invention uses the exponential function +.>To enhance this gap so that the loss function +.>The calculated loss value is large, the weight and bias in the voice recognition unit are reduced greatly, and the training time is shortened.
The technical scheme of the embodiment of the invention has at least the following advantages and beneficial effects: according to the invention, the voice signal of the physical examination personnel is collected through the voice collecting and denoising unit, the voice signal is denoised, the recognition precision is improved, the voice recognition unit is used for recognizing the denoised voice signal, the requirement of the physical examination personnel is obtained, the requirement keywords are extracted, the requirement keywords of the physical examination personnel are matched with the descriptions of each physical examination item, the successfully matched physical examination items and the corresponding physical examination process are pushed to the physical examination personnel, the workload of medical care work is reduced, and the queuing time of the physical examination personnel is further reduced.
Drawings
FIG. 1 is a system block diagram of a physical examination item recommendation system based on personal needs;
FIG. 2 is a block diagram of a speech recognition unit;
fig. 3 is a block diagram of the residual module.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. The components of the embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.
As shown in fig. 1, a physical examination item recommendation system based on personal needs includes: the device comprises a voice acquisition denoising unit, a voice recognition unit, a keyword extraction unit, a keyword matching unit and a physical examination item pushing unit;
the voice acquisition denoising unit is used for acquiring voice signals of physical examination personnel and performing wavelet transformation denoising on the voice signals to obtain denoised voice signals; the voice recognition unit is used for recognizing the denoising voice signal to obtain text information; the keyword extraction unit is used for extracting physical examination keywords from the text information; the keyword matching unit is used for matching the physical examination keywords with the descriptions of each physical examination item and selecting the physical examination item successfully matched; the physical examination item pushing unit is used for pushing the physical examination items successfully matched with the corresponding physical examination processes.
In this embodiment, the keyword extraction unit may store the keywords in the physical examination description in the memory, so as to match the keywords in the memory with the text information, the successfully matched keywords are the required keywords, and then match the keywords with the text description in the physical examination item, and when the keywords are successfully matched, the corresponding physical examination item is the required physical examination item.
The voice acquisition denoising unit comprises: a wavelet decomposition subunit, a wavelet coefficient selection subunit, and a reconstruction subunit;
the wavelet decomposition subunit is used for performing wavelet decomposition on the voice signal to obtain wavelet coefficients;
the wavelet coefficient selection subunit is used for updating the wavelet coefficient according to the threshold value to obtain an updated wavelet coefficient;
the reconstruction subunit is used for carrying out reconstruction processing on the updated wavelet coefficient to obtain a denoising voice signal.
The expression of the updated wavelet coefficients is:
wherein ,for the updated->Layer wavelet coefficients, < >>For the%>Layer wavelet coefficients, < >>To update the coefficients.
The functions of the existing updated wavelet coefficients are segment functions, and the functions are discontinuous at the joint points of the segment functions, so that the precision of the updated wavelet coefficients is not high, therefore, the invention utilizes the continuous smoothing function with the function value between-1 and 1A new expression for updating the wavelet coefficients is constructed, so that the whole definition domain range is smooth, the wavelet coefficients can be precisely removed, and the denoising precision is improved.
The update coefficientThe expression of (2) is:
wherein ,is->Threshold value of layer wavelet coefficient, < >>For the%>Layer wavelet coefficients.
The invention updates the coefficientFollowing->Layer wavelet coefficients->And threshold->The adaptive change ensures that the function of the wavelet coefficient has good transition near the threshold value.
Said firstThreshold value of layer wavelet coefficient->The expression of (2) is:
wherein ,is a proportional coefficient->Is->The>Wavelet coefficient value,/">Is the number of wavelet coefficient values, +.>For the length of the speech signal, < >>For wavelet decomposition scale, +.>As a logarithmic function.
The invention utilizes the firstThe average value of wavelet coefficient values in the wavelet coefficients of the layers is used for estimating the threshold value, and the proportional coefficient is set for adjusting the threshold value.
As shown in fig. 2, the voice recognition unit includes: the system comprises a convolution module, a residual error module, a first LSTM module, a second LSTM module, an attention module, a CNN network and a CTC classifier;
the input end of the convolution module is connected with the input end of the first LSTM module and is used as the input end of the voice recognition unit; the output end of the convolution module is connected with the input end of the residual error module; the input end of the second LSTM module is connected with the output end of the first LSTM module; the input end of the attention module is respectively connected with the output end of the residual error module and the output end of the second LSTM module, and the output end of the attention module is connected with the input end of the CNN network; the input end of the CTC classifier is connected with the output end of the CNN network, and the output end of the CTC classifier is used as the output end of the voice recognition unit.
The invention utilizes two paths to extract the characteristics of the denoising voice signal respectively, enriches the characteristic quantity, utilizes the LSTM module to have time memory, can better consider the history characteristics, the residual error module fuses the deep level and the shallow level sub-characteristics, the output of the LSTM module and the output of the residual error module are weighted and fused in the attention module, the processing is carried out according to the significance degree of each characteristic, the weight is applied to each characteristic in a self-adaption manner, the weighted and fused characteristics are input into the CNN network to carry out the deep extraction of the characteristics, and then the text information is output through the CTC classifier.
As shown in fig. 3, the residual module includes: the system comprises a first convolution sub-module, a second convolution sub-module, a third convolution sub-module, an adder and a multiplier;
the input end of the first convolution sub-module is respectively connected with the input end of the third convolution sub-module and the first input end of the multiplier, and the output end of the first convolution sub-module is connected with the input end of the second convolution sub-module; the first input end of the adder is connected with the output end of the second convolution sub-module, the second input end of the adder is connected with the output end of the third convolution sub-module, and the output end of the adder is connected with the second input end of the multiplier; the output end of the multiplier is used as the output end of the residual error module.
According to the invention, the features processed by the two convolution sub-modules and the features processed by one convolution sub-module are added through the adder, so that the information quantity is improved, and then the multiplier is used for multiplying the input features, so that the problem of gradient disappearance can be solved, and the shallow features can be fused.
In the present invention, the convolution sub-module and the convolution module each include: convolution layer, reLU layer and BN layer.
The expression of the attention module is:
wherein ,for the output of the attention module, +.>For Concat splice operation, < >>Is->Weight of->For the output of the residual block,/>Is->Weight of->Is the output of the second LSTM module.
The saidWeight of +.>The expression of (2) is:
wherein ,as an exponential function based on natural constants, < +.>To activate the function +.>For global pooling processing,/->Is the output of the residual error module;
the saidWeight of +.>The expression of (2) is:
wherein ,is the output of the second LSTM module.
In the invention respectively to and />Different weights are given, and splicing is carried out, so that feature fusion is realized, and the method is characterized in that +_> and />According to-> and />The self situation is calculated, and the self-adaption attention degree of the remarkable characteristics is improved.
The loss function of the voice recognition unit is as follows:
wherein ,for loss function->Is->Status function at sub-training>The true class in the denoised speech signal samples is equal to the class +.>When 1 is taken, if not, 0 is taken, +.>As an exponential function based on natural constants, < +.>Is->Predictive probability of speech recognition unit during secondary training, < >>For the current training times, +.>For a small scale adjacent to the training times, +.>For the total number of adjacent exercises>Is the number of categories.
In the present inventionWhen equal to 1>The closer to 0, the more the prediction differs from the tag, so the present invention uses the exponential function +.>To enhance this gap so that the loss function +.>The calculated loss value is large, the weight and bias in the voice recognition unit are reduced greatly, and the training time is shortened.
In the present invention,for marking the current training times>For marking the number of adjacent exercises, selecting adjacent exercises>And (5) comprehensively evaluating the training condition according to the secondary condition.
The technical scheme of the embodiment of the invention has at least the following advantages and beneficial effects: according to the invention, the voice signal of the physical examination personnel is collected through the voice collecting and denoising unit, the voice signal is denoised, the recognition precision is improved, the voice recognition unit is used for recognizing the denoised voice signal, the requirement of the physical examination personnel is obtained, the requirement keywords are extracted, the requirement keywords of the physical examination personnel are matched with the descriptions of each physical examination item, the successfully matched physical examination items and the corresponding physical examination process are pushed to the physical examination personnel, the workload of medical care work is reduced, and the queuing time of the physical examination personnel is further reduced.
The above is only a preferred embodiment of the present invention, and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (6)

1. A physical examination item recommendation system based on personal needs, comprising: the device comprises a voice acquisition denoising unit, a voice recognition unit, a keyword extraction unit, a keyword matching unit and a physical examination item pushing unit;
the voice acquisition denoising unit is used for acquiring voice signals of physical examination personnel and performing wavelet transformation denoising on the voice signals to obtain denoised voice signals; the voice recognition unit is used for recognizing the denoising voice signal to obtain text information; the keyword extraction unit is used for extracting physical examination keywords from the text information; the keyword matching unit is used for matching the physical examination keywords with the descriptions of each physical examination item and selecting the physical examination item successfully matched; the physical examination item pushing unit is used for pushing the physical examination items successfully matched with the corresponding physical examination processes;
the voice recognition unit includes: the system comprises a convolution module, a residual error module, a first LSTM module, a second LSTM module, an attention module, a CNN network and a CTC classifier;
the input end of the convolution module is connected with the input end of the first LSTM module and is used as the input end of the voice recognition unit; the output end of the convolution module is connected with the input end of the residual error module; the input end of the second LSTM module is connected with the output end of the first LSTM module; the input end of the attention module is respectively connected with the output end of the residual error module and the output end of the second LSTM module, and the output end of the attention module is connected with the input end of the CNN network; the input end of the CTC classifier is connected with the output end of the CNN network, and the output end of the CTC classifier is used as the output end of the voice recognition unit;
the residual error module comprises: the system comprises a first convolution sub-module, a second convolution sub-module, a third convolution sub-module, an adder and a multiplier;
the input end of the first convolution sub-module is respectively connected with the input end of the third convolution sub-module and the first input end of the multiplier, and the output end of the first convolution sub-module is connected with the input end of the second convolution sub-module; the first input end of the adder is connected with the output end of the second convolution sub-module, the second input end of the adder is connected with the output end of the third convolution sub-module, and the output end of the adder is connected with the second input end of the multiplier; the output end of the multiplier is used as the output end of the residual error module;
the expression of the attention module is:
wherein ,for the output of the attention module, +.>For Concat splice operation, < >>Is->Weight of->For the output of the residual block,/>Is->Weight of->Is the output of the second LSTM module;
the saidWeight of +.>The expression of (2) is:
wherein ,as an exponential function based on natural constants, < +.>To activate the function +.>For global pooling processing,/->Is the output of the residual error module;
the saidWeight of +.>The expression of (2) is:
wherein ,is the output of the second LSTM module.
2. The personal demand based physical examination item recommendation system of claim 1, wherein the voice acquisition denoising unit comprises: a wavelet decomposition subunit, a wavelet coefficient selection subunit, and a reconstruction subunit;
the wavelet decomposition subunit is used for performing wavelet decomposition on the voice signal to obtain wavelet coefficients;
the wavelet coefficient selection subunit is used for updating the wavelet coefficient according to the threshold value to obtain an updated wavelet coefficient;
the reconstruction subunit is used for carrying out reconstruction processing on the updated wavelet coefficient to obtain a denoising voice signal.
3. The personal demand based physical examination item recommendation system of claim 2, wherein the expression of the updated wavelet coefficients is:
wherein ,for the updated->Layer wavelet coefficients, < >>For the%>Layer wavelet coefficients, < >>To update the coefficients.
4. The personal demand based physical examination item recommendation system of claim 3, wherein the update coefficientsThe expression of (2) is:
wherein ,is->Threshold value of layer wavelet coefficient, < >>For the%>Layer wavelet coefficients.
5. The personal demand based physical examination item recommendation system of claim 4, wherein the firstThreshold value of layer wavelet coefficient->The expression of (2) is:
wherein ,is a proportional coefficient->Is->The>Wavelet coefficient value,/">Is the number of wavelet coefficient values, +.>For the length of the speech signal, < >>For wavelet decomposition scale, +.>As a logarithmic function.
6. The personal demand based physical examination item recommendation system of claim 1 wherein said voice recognition unit has a loss function of:
wherein ,for loss function->Is->Status function at sub-training>The true class in the denoised speech signal samples is equal to the class +.>When 1 is taken, if not, 0 is taken, +.>As an exponential function based on natural constants, < +.>Is the firstPredictive probability of speech recognition unit during secondary training, < >>For the current training times, +.>For a small scale adjacent to the training times, +.>For the total number of adjacent exercises>Is the number of categories.
CN202311035553.6A 2023-08-17 2023-08-17 Physical examination project recommendation system based on personal demands Active CN116759061B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311035553.6A CN116759061B (en) 2023-08-17 2023-08-17 Physical examination project recommendation system based on personal demands

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311035553.6A CN116759061B (en) 2023-08-17 2023-08-17 Physical examination project recommendation system based on personal demands

Publications (2)

Publication Number Publication Date
CN116759061A CN116759061A (en) 2023-09-15
CN116759061B true CN116759061B (en) 2023-10-27

Family

ID=87957477

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311035553.6A Active CN116759061B (en) 2023-08-17 2023-08-17 Physical examination project recommendation system based on personal demands

Country Status (1)

Country Link
CN (1) CN116759061B (en)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106355321A (en) * 2016-08-29 2017-01-25 北京红辣椒信息科技有限公司 Physical examination scheduling device and scheduling method capable of selecting items by oneself
CN110189749A (en) * 2019-06-06 2019-08-30 四川大学 Voice keyword automatic identifying method
CN110246490A (en) * 2019-06-26 2019-09-17 合肥讯飞数码科技有限公司 Voice keyword detection method and relevant apparatus
CN112289309A (en) * 2020-10-30 2021-01-29 西安工程大学 Robot voice control method based on deep learning
CN112330713A (en) * 2020-11-26 2021-02-05 南京工程学院 Method for improving speech comprehension degree of severe hearing impaired patient based on lip language recognition
CN113724860A (en) * 2021-08-31 2021-11-30 平安国际智慧城市科技股份有限公司 Medical examination recommendation method, device, equipment and medium based on artificial intelligence
CN114372201A (en) * 2022-01-11 2022-04-19 平安科技（深圳）有限公司 Physical examination information intelligent recommendation method and system, storage medium and computing equipment
CN115942899A (en) * 2020-11-10 2023-04-07 索尼集团公司 Medical examination of the human body using tactile sensation

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP6703460B2 (en) * 2016-08-25 2020-06-03 本田技研工業株式会社 Audio processing device, audio processing method, and audio processing program

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106355321A (en) * 2016-08-29 2017-01-25 北京红辣椒信息科技有限公司 Physical examination scheduling device and scheduling method capable of selecting items by oneself
CN110189749A (en) * 2019-06-06 2019-08-30 四川大学 Voice keyword automatic identifying method
CN110246490A (en) * 2019-06-26 2019-09-17 合肥讯飞数码科技有限公司 Voice keyword detection method and relevant apparatus
CN112289309A (en) * 2020-10-30 2021-01-29 西安工程大学 Robot voice control method based on deep learning
CN115942899A (en) * 2020-11-10 2023-04-07 索尼集团公司 Medical examination of the human body using tactile sensation
CN112330713A (en) * 2020-11-26 2021-02-05 南京工程学院 Method for improving speech comprehension degree of severe hearing impaired patient based on lip language recognition
CN113724860A (en) * 2021-08-31 2021-11-30 平安国际智慧城市科技股份有限公司 Medical examination recommendation method, device, equipment and medium based on artificial intelligence
CN114372201A (en) * 2022-01-11 2022-04-19 平安科技（深圳）有限公司 Physical examination information intelligent recommendation method and system, storage medium and computing equipment

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
基于小波神经网络的嵌入式语音识别系统;李卫;宋弘;姜天华;;通信技术(06);第213-215+218页 *
多模态深度学习综述;刘建伟;丁熙浩;罗雄麟;;计算机应用研究(06);第1601-1614页 *
王仕奎.《随机信号分析理论与实践》.南京：东南大学出版社,2016,202-203. *
简化体格检查188条在诊断学教学中的应用;雷红, 张颖, 张克俭, 占建珍, 罗宁, 陈惠芳;西北医学教育(02);第97-99页 *

Also Published As

Publication number Publication date
CN116759061A (en) 2023-09-15

Similar Documents

Publication Publication Date Title
CN111192680B (en) Intelligent auxiliary diagnosis method based on deep learning and collective classification
CN109190110B (en) Named entity recognition model training method and system and electronic equipment
CN107391614A (en) A kind of Chinese question and answer matching process based on WMD
JP2012118977A (en) Method and system for machine-learning based optimization and customization of document similarity calculation
CN112559684A (en) Keyword extraction and information retrieval method
CN112507118A (en) Information classification and extraction method and device and electronic equipment
CN111144068A (en) Similar arbitration case recommendation method and device
CN114359974A (en) Human body posture detection method and device and storage medium
CN115688920A (en) Knowledge extraction method, model training method, device, equipment and medium
CN113496122A (en) Named entity identification method, device, equipment and medium
CN112818227B (en) Content recommendation method and device, electronic equipment and storage medium
CN116467461A (en) Data processing method, device, equipment and medium applied to power distribution network
CN113722507B (en) Hospitalization cost prediction method and device based on knowledge graph and computer equipment
CN113827240B (en) Emotion classification method, training device and training equipment for emotion classification model
CN115248890B (en) User interest portrait generation method and device, electronic equipment and storage medium
CN115457329A (en) Training method of image classification model, image classification method and device
CN116759061B (en) Physical examination project recommendation system based on personal demands
CN113657248A (en) Training method and device for face recognition model and computer program product
CN117493595A (en) Image searching method, device, equipment and medium based on large model
CN117057350A (en) Chinese electronic medical record named entity recognition method and system
CN115391536A (en) Enterprise public opinion identification method, device, equipment and storage medium
CN115761219A (en) Image detection method, device, equipment and medium for disabled elderly
CN114664436A (en) First aid auxiliary system based on intelligent agent decision
CN113808619A (en) Voice emotion recognition method and device and electronic equipment
CN114443864A (en) Cross-modal data matching method and device and computer program product

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant