Parkinson's disease early diagnosis system based on intelligent voice recognition analysis

Abstract

The invention discloses a Parkinson's disease early diagnosis system based on intelligent speech recognition analysis, and belongs to the technical field of Parkinson's disease, and the system comprises a speech collection module which is used for collecting user speech real-time data; the voice processing module is used for processing user voice real-time data; the model training module is used for constructing a Parkinson's disease early diagnosis model; the test optimization module is used for performing test optimization on the Parkinson's disease early diagnosis model; the analysis and diagnosis module is used for carrying out early diagnosis on the Parkinson's disease of the user; and the display management module is used for carrying out early warning on the Parkinson's disease early diagnosis of the user. According to the invention, the problem of poor early diagnosis effect of Parkinson's disease caused by incapability of early diagnosis of Parkinson's disease based on intelligent speech recognition analysis in the prior art is solved. According to the invention, the Parkinson's disease can be diagnosed in an early stage based on intelligent voice recognition analysis, the Parkinson's disease can be found and treated in the early stage, and the early diagnosis effect of the Parkinson's disease can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of Parkinson's disease, and in particular to an early diagnosis system for Parkinson's disease based on intelligent speech recognition analysis. Background Art

[0002] Parkinson's disease, also often referred to as "tremor paralysis", is a neurodegenerative disease. Its main cause is the degeneration and death of dopaminergic neurons in the substantia nigra. It may be related to multiple factors such as genetics, environmental factors and aging of the nervous system. It is generally recognized that aging is the most important factor in the occurrence of Parkinson's disease. The disease has a significant high incidence in the elderly, and the incidence rate in men is slightly higher than that in women. Its symptoms usually show obvious clinical manifestations only in the later stages of the disease. Therefore, early diagnosis of Parkinson's disease is particularly important.

[0003] Chinese patent publication number CN117222355A discloses a method for treating, preventing or ameliorating Parkinson's disease with putinizumab, wherein a mobile device is provided to a patient, the mobile device is programmed to receive and transmit data obtained to measure the patient's passive and / or active movement; (b) data transmitted from the mobile device is collected; and (c) the data obtained from the patient is compared with control data to assess the presence or extent of motor deficits in an individual, and / or the data obtained from the patient is monitored for a period of time sufficient to identify changes in the patient's active or passive motor function; however, the patent has the following defects: The existing methods cannot perform early diagnosis of Parkinson's disease based on intelligent speech recognition analysis, and cannot achieve early detection and treatment of Parkinson's disease, resulting in poor results in early diagnosis of Parkinson's disease. Summary of the invention

[0004] The purpose of the present invention is to provide an early diagnosis system for Parkinson's disease based on intelligent speech recognition analysis, which can perform early diagnosis of Parkinson's disease based on intelligent speech recognition analysis, can achieve early detection and treatment of Parkinson's disease, can improve the early diagnosis effect of Parkinson's disease, and solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions: The Parkinson's disease early diagnosis system based on intelligent speech recognition analysis includes: The voice collection module is used to monitor the user's recording in real time and capture the user's voice signal, and collect the user's real-time voice data; The voice processing module is used to perform noise reduction, enhancement, word segmentation and feature extraction on the real-time user voice data to determine the user voice feature data; A model training module is used to build an early diagnosis model for Parkinson's disease based on intelligent speech recognition analysis according to the early diagnosis needs of Parkinson's disease based on intelligent speech recognition analysis; A test optimization module is used to test and optimize the Parkinson's disease early diagnosis model based on intelligent speech recognition analysis to determine the optimal Parkinson's disease early diagnosis model based on intelligent speech recognition analysis; An analysis and diagnosis module is used to mine and analyze the user's voice feature data based on the optimal Parkinson's disease early diagnosis model based on intelligent voice recognition analysis, to perform early diagnosis of Parkinson's disease on the user, and to determine the user's Parkinson's disease early diagnosis result based on intelligent voice recognition analysis; The display management module is used to warn users of early diagnosis of Parkinson's disease and to display and manage the results of early diagnosis of Parkinson's disease for users.

[0006] Preferably, the voice collection module includes: A voice capture unit, used to monitor user recordings in real time and capture user voice signals; According to the needs of early diagnosis of Parkinson's disease based on intelligent speech recognition analysis, real-time monitoring of user recordings is carried out to capture user voice signals; Voice collection unit, used to collect real-time data of user voice; Acquire user voice signals, collect user voice signals, and determine user voice real-time data.

[0007] Preferably, the speech processing module comprises: A speech noise reduction unit is used to perform noise reduction processing on real-time user speech data; Get real-time user voice data; Based on the median filter, the real-time user voice data is de-noised; Remove noise from real-time user voice data; A speech enhancement unit is used to enhance the real-time data of user speech; Obtain real-time user voice data after noise reduction; The speech enhancement algorithm based on Kalman filtering can enhance the real-time user speech data after noise reduction, suppress and reduce noise interference; Real-time user speech data valuable for early diagnosis of Parkinson's disease is extracted from the noise background.

[0008] Preferably, the voice processing module further includes: A first volume data extraction module is used to extract the volume of each frame of audio in the real-time voice data; A first comparison module, used for comparing the volume of each frame of audio with a preset volume threshold; An audio frame number extraction module is used to extract the number of audio frames whose volume is lower than a preset volume threshold; A second comparison module, used for comparing the number of audio frames with a preset frame number threshold; A second volume data extraction module is used to retrieve the volume value corresponding to the audio below the preset volume threshold when the number of audio frames exceeds the preset frame number threshold; A volume index value acquisition module, used to obtain a volume index value using the volume value corresponding to the audio that is lower than the preset volume threshold; The volume index value is obtained by the following formula:

[0009] Wherein, K represents the volume index value; n represents the number of audio frames below the preset volume threshold; F i Indicates the volume value of the audio of the i-th frame that is lower than the preset volume threshold; F y Indicates the preset volume threshold; F max Indicates the maximum volume of the audio that is lower than the preset volume threshold; F min Indicates the minimum volume value of the audio that is lower than the preset volume threshold; F m represents the preset minimum allowed volume value; f represents the adjustment coefficient, and the adjustment coefficient is obtained by the following formula:

[0010] Wherein, f represents the adjustment coefficient; n represents the number of audio frames below the preset volume threshold; F i Indicates the volume value of the audio of the i-th frame that is lower than the preset volume threshold; F y Indicates the preset volume threshold; F m Indicates the preset minimum allowed volume value; A third comparison module, used for comparing the volume index value with a preset index value threshold; The volume adjustment module is used to adjust the volume of the audio when the volume index value is lower than a preset index value threshold.

[0011] Preferably, the volume adjustment module includes: A first audio data acquisition module is used to extract the audio volume that exceeds the preset volume threshold and has the closest time distance before and after each frame of audio that is lower than the preset volume threshold as the first audio data; A second audio data acquisition module, used to extract the volume of each frame of audio below the preset volume threshold as second audio data; A volume index value extraction module, used to extract the volume index value; A volume enhancement adjustment module, configured to perform volume enhancement adjustment on each frame of audio that is lower than the preset volume threshold using the first audio data, the second audio data and the volume index value; The volume of each frame of audio that is lower than the preset volume threshold after enhancement and adjustment is obtained by the following formula:

[0012] Wherein, F represents the volume of each frame of audio that is lower than the preset volume threshold after adjustment; F0 represents the volume of each frame of audio that is lower than the preset volume threshold; F x represents the volume of the audio frame that is closer to the preset volume threshold than the audio frame that is closer to the preset volume threshold; K represents the volume index value; f represents the adjustment coefficient; f represents the adjustment coefficient; y represents the adjustment factor, and the adjustment factor is obtained by the following formula: Among them, f y represents the adjustment factor; n represents the number of audio frames below the preset volume threshold; F i Indicates the volume value of the audio of the i-th frame that is lower than the preset volume threshold; F y Indicates the preset volume threshold.

[0013] Preferably, the voice processing module further includes: A speech segmentation unit is used to perform word segmentation processing on real-time user speech data; Obtain enhanced real-time user voice data; Based on natural language processing technology, the real-time user voice data is segmented; Cut the continuous real-time user voice data into a series of small semantic units that are easy to analyze; A feature extraction unit, used to extract features from real-time user voice data; Get real-time user voice data after word segmentation; Based on the Fourier transform method, the features of speech pronunciation, intonation, and speaking speed that are valuable for the early diagnosis of Parkinson's disease are extracted from the real-time user speech data after word segmentation, and the user speech feature data is determined.

[0014] Preferably, the model training module includes: A data collection unit, used to collect user voice history data; According to the needs of early diagnosis of Parkinson's disease based on intelligent speech recognition analysis, user voice history data is collected and divided; Determine the Parkinson's disease early diagnosis training set and Parkinson's disease early diagnosis test set; A model building unit, used to build an early diagnosis model for Parkinson's disease based on intelligent speech recognition analysis; According to the demand for early diagnosis of Parkinson's disease based on intelligent speech recognition analysis, a convolutional neural network model framework suitable for early diagnosis of Parkinson's disease is selected; Based on the Parkinson's disease early diagnosis training set, the selected convolutional neural network model framework suitable for the early diagnosis of Parkinson's disease is trained; An early diagnosis model for Parkinson's disease based on intelligent speech recognition analysis was determined.

[0015] Preferably, the test optimization module includes: A test and evaluation unit, used to test and evaluate the Parkinson's disease early diagnosis model based on intelligent speech recognition analysis; Obtain an early diagnosis model for Parkinson's disease based on intelligent speech recognition analysis; Based on the Parkinson's disease early diagnosis test set, the performance test and evaluation of the Parkinson's disease early diagnosis model based on intelligent speech recognition analysis was conducted; Determine the performance test evaluation results based on the Parkinson's disease early diagnosis model; An adjustment and optimization unit, used to adjust and optimize the Parkinson's disease early diagnosis model based on intelligent speech recognition analysis; Obtain performance test evaluation results based on the Parkinson's disease early diagnosis model; According to the performance test evaluation results of the Parkinson's disease early diagnosis model, the Parkinson's disease early diagnosis model based on intelligent speech recognition analysis is mined and analyzed; Determine the parameter adjustment optimization scheme based on the early diagnosis model of Parkinson's disease; According to the parameter adjustment optimization scheme based on the Parkinson's disease early diagnosis model, the parameters of the Parkinson's disease early diagnosis model based on intelligent speech recognition analysis are adjusted and optimized; Determine the optimal early diagnosis model for Parkinson's disease based on intelligent speech recognition analysis.

[0016] Preferably, the analysis and diagnosis module includes: A data extraction unit, used to extract user voice feature data; An analysis and diagnosis unit, used for early diagnosis of Parkinson's disease for users; Obtain the best early diagnosis model for Parkinson's disease based on intelligent speech recognition analysis; Input the user's voice feature data into the optimal Parkinson's disease early diagnosis model based on intelligent speech recognition analysis; Based on the optimal Parkinson's disease early diagnosis model based on intelligent speech recognition analysis, the user's speech feature data is mined and analyzed to perform early diagnosis of Parkinson's disease for the user; Determine the early diagnosis results of users' Parkinson's disease based on intelligent speech recognition analysis.

[0017] Preferably, the display management module includes: Diagnosis and early warning unit, used to warn users of early diagnosis of Parkinson's disease; Obtain early diagnosis results of users’ Parkinson’s disease based on intelligent speech recognition analysis; According to the early diagnosis results of Parkinson's disease of users based on intelligent speech recognition analysis, timely warning of early diagnosis of Parkinson's disease of users can be given, so that users can receive follow-up treatment in time; A display management unit, used to display and manage the user's Parkinson's disease early diagnosis results; Obtain early diagnosis results of users’ Parkinson’s disease based on intelligent speech recognition analysis; The early diagnosis results of the user's Parkinson's disease based on the intelligent speech recognition analysis are displayed to the user in a visual form, and the early diagnosis results of the user's Parkinson's disease based on the intelligent speech recognition analysis are stored and managed.

[0018] Compared with the prior art, the present invention has the following beneficial effects: The present invention collects real-time data of user voice by real-time monitoring of user recordings and capturing user voice signals, determines user voice feature data by performing noise reduction, enhancement, word segmentation and feature extraction processing on the real-time user voice data, constructs a Parkinson's disease early diagnosis model based on intelligent voice recognition analysis according to the demand for early diagnosis of Parkinson's disease based on intelligent voice recognition analysis, tests and optimizes the Parkinson's disease early diagnosis model based on intelligent voice recognition analysis, determines the optimal Parkinson's disease early diagnosis model based on intelligent voice recognition analysis, mines and analyzes user voice feature data based on the optimal Parkinson's disease early diagnosis model based on intelligent voice recognition analysis, performs early diagnosis of Parkinson's disease on users, determines the results of early diagnosis of Parkinson's disease on users based on intelligent voice recognition analysis, and performs early warning on the early diagnosis of Parkinson's disease on users, displays and manages the results of early diagnosis of Parkinson's disease on users, can perform early diagnosis of Parkinson's disease based on intelligent voice recognition analysis, can achieve early discovery and treatment of Parkinson's disease, and can improve the effect of early diagnosis of Parkinson's disease. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a schematic diagram of the structure of the Parkinson's disease early diagnosis system based on intelligent speech recognition analysis of the present invention; Figure 2This is a structural block diagram of the Parkinson's disease early diagnosis system based on intelligent speech recognition analysis of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] In order to solve the problem that the existing intelligent speech recognition analysis cannot be used to diagnose Parkinson's disease early, and the early detection and treatment of Parkinson's disease cannot be achieved, resulting in poor results in the early diagnosis of Parkinson's disease, please refer to Figure 1-2 , this embodiment provides the following technical solutions: The Parkinson's disease early diagnosis system based on intelligent speech recognition analysis includes: a speech collection module, a speech processing module, a model training module, a test optimization module, an analysis and diagnosis module and a display management module, wherein the speech collection module, the speech processing module, the model training module, the test optimization module, the analysis and diagnosis module and the display management module interact and communicate with each other.

[0022] The voice collection module is used to monitor the user's voice recording in real time and capture the user's voice signal to collect the user's voice real-time data; In this embodiment, the voice collection module includes: A voice capture unit, used to monitor user recordings in real time and capture user voice signals; According to the needs of early diagnosis of Parkinson's disease based on intelligent speech recognition analysis, real-time monitoring of user recordings is carried out to capture user voice signals; Voice collection unit, used to collect real-time data of user voice; Acquire user voice signals, collect user voice signals, and determine user voice real-time data.

[0023] The voice processing module is used to perform noise reduction, enhancement, word segmentation and feature extraction on the real-time user voice data to determine the user voice feature data; In this embodiment, the speech processing module includes: A speech noise reduction unit is used to perform noise reduction processing on real-time user speech data; Get real-time user voice data; Based on the median filter, the real-time user voice data is de-noised; Remove noise from real-time user voice data; A speech enhancement unit is used to enhance the real-time data of user speech; Obtain real-time user voice data after noise reduction; The speech enhancement algorithm based on Kalman filtering can enhance the real-time user speech data after noise reduction, suppress and reduce noise interference; Extracting real-time user voice data from a noisy background, which is valuable for early diagnosis of Parkinson's disease; Specifically, the voice processing module further includes: A first volume data extraction module is used to extract the volume of each frame of audio in the real-time voice data; A first comparison module, used for comparing the volume of each frame of audio with a preset volume threshold; An audio frame number extraction module is used to extract the number of audio frames whose volume is lower than a preset volume threshold; A second comparison module, used for comparing the number of audio frames with a preset frame number threshold; A second volume data extraction module is used to retrieve the volume value corresponding to the audio below the preset volume threshold when the number of audio frames exceeds the preset frame number threshold; A volume index value acquisition module, used to obtain a volume index value using the volume value corresponding to the audio that is lower than the preset volume threshold; The volume index value is obtained by the following formula: Wherein, K represents the volume index value; n represents the number of audio frames below the preset volume threshold; F i Indicates the volume value of the audio of the i-th frame that is lower than the preset volume threshold; F y Indicates the preset volume threshold; F max Indicates the maximum volume of the audio that is lower than the preset volume threshold; F min Indicates the minimum volume value of the audio that is lower than the preset volume threshold; F m represents the preset minimum allowed volume value; f represents the adjustment coefficient, and the adjustment coefficient is obtained by the following formula: Wherein, f represents the adjustment coefficient; n represents the number of audio frames below the preset volume threshold; F i Indicates the volume value of the audio of the i-th frame that is lower than the preset volume threshold; F y Indicates the preset volume threshold; F m Indicates the preset minimum allowed volume value; A third comparison module, used for comparing the volume index value with a preset index value threshold; The volume adjustment module is used to adjust the volume of the audio when the volume index value is lower than a preset index value threshold.

[0024] The technical effect of the above technical solution is: through the first volume data extraction module and the first comparison module, the volume of each frame of audio can be accurately extracted and compared with the preset volume threshold. This helps to identify which parts have low volume and may need to be adjusted. The audio frame number extraction module and the second comparison module are used in combination to count and determine whether the number of audio frames with volume below the preset threshold exceeds a certain threshold. This helps to identify whether there is a continuous problem of low volume, not just occasional volume fluctuations. When a volume problem is detected (that is, the number of audio frames exceeds the preset frame number threshold), the second volume data extraction module will further extract the specific volume values ​​of these low-volume audios. This provides detailed data support for subsequent volume index calculations. The volume index value acquisition module uses the extracted volume data to calculate the volume index value through a specific formula. This value comprehensively considers the number of frames of low-volume audio, the distribution of volume values, and the preset volume threshold and the minimum allowable volume value. This makes the volume evaluation more comprehensive and scientific. The introduction of the adjustment coefficient makes the calculation of the volume index more flexible. It is dynamically adjusted according to the specific situation of the low-volume audio and the preset minimum allowable volume value, which helps to more accurately reflect the volume status of the audio. Through the third comparison module and the volume adjustment module, when the volume index value is lower than the preset index value threshold, the system will adjust the volume of the audio. This helps to improve the auditory effect of the audio, making it clearer and easier to understand. This technical solution realizes comprehensive analysis and processing of real-time voice data through a series of sophisticated module designs. This helps to improve the overall performance of the voice processing system, enabling it to better adapt to various complex voice environments.

[0025] In summary, this technical solution achieves comprehensive optimization and processing of real-time voice data through precise volume analysis, intelligent volume problem detection, detailed volume data extraction, scientific volume index calculation, flexible adjustment coefficient calculation and effective volume adjustment. This helps to improve the auditory effect of voice and improve user experience.

[0026] Specifically, the volume adjustment module includes: A first audio data acquisition module is used to extract the audio volume that exceeds the preset volume threshold and has the closest time distance before and after each frame of audio that is lower than the preset volume threshold as the first audio data; A second audio data acquisition module, used to extract the volume of each frame of audio below the preset volume threshold as second audio data; A volume index value extraction module, used to extract the volume index value; A volume enhancement adjustment module, configured to perform volume enhancement adjustment on each frame of audio that is lower than the preset volume threshold using the first audio data, the second audio data and the volume index value; The volume of each frame of audio that is lower than the preset volume threshold after enhancement and adjustment is obtained by the following formula:

[0027] Wherein, F represents the volume of each frame of audio that is lower than the preset volume threshold after adjustment; F0 represents the volume of each frame of audio that is lower than the preset volume threshold; F x represents the volume of the audio frame that is closer to the preset volume threshold than the audio frame that is closer to the preset volume threshold; K represents the volume index value; f represents the adjustment coefficient; f represents the adjustment coefficient; y represents the adjustment factor, and the adjustment factor is obtained by the following formula: Among them, f y represents the adjustment factor; n represents the number of audio frames below the preset volume threshold; F i Indicates the volume value of the audio of the i-th frame that is lower than the preset volume threshold; F y Indicates the preset volume threshold.

[0028] The technical effect of the above technical solution is as follows: the first audio data acquisition module and the second audio data acquisition module respectively extract important volume data related to the low volume audio frame. The first audio data is the audio volume that exceeds the preset volume threshold value with the closest time distance before and after the low volume frame, which provides a reference benchmark for volume enhancement; the second audio data is the volume of the low volume frame itself, which is the object that needs to be adjusted directly. The volume enhancement adjustment module uses the first audio data (i.e., the volume of the surrounding high volume frames), the second audio data (i.e., the volume of the low volume frame) and the volume index value to comprehensively determine the enhancement degree of each frame of low volume audio. This strategy not only considers the volume level of the low volume frame itself, but also considers its relative relationship with the surrounding high volume frames, thereby achieving a more reasonable and natural volume enhancement effect. The introduction of the adjustment coefficient and the adjustment factor makes the volume enhancement adjustment more flexible and controllable. The adjustment coefficient may be calculated according to a specific algorithm or rule to adjust the amplitude of the volume enhancement; while the adjustment factor is dynamically calculated according to the relationship between the volume value of the low volume frame and the preset volume threshold, further refining the volume enhancement strategy. The volume adjustment formula provided takes into account a variety of factors, including the volume of the low-volume frame itself, the volume of the surrounding high-volume frames, the volume index value, the adjustment coefficient and adjustment factor, etc. This allows the adjusted volume to meet the preset volume threshold requirements while maintaining the naturalness and coherence of the audio. Through the implementation of the above technical solution, the auditory effect of the low-volume part of the audio can be significantly improved. The enhanced low-volume frame will be clearer and easier to understand, thereby improving the auditory quality of the overall audio. The optimized processing of the volume adjustment module makes the audio content smoother and more comfortable during playback, avoiding hearing impairment caused by too low volume. This helps to improve user satisfaction and experience with audio content.

[0029] In summary, this technical solution achieves effective enhancement and adjustment of low-volume audio frames through precise volume data acquisition, scientific volume enhancement strategy, flexible adjustment coefficient and adjustment factor, and detailed volume adjustment formula, which not only improves the auditory quality of the audio, but also enhances the user's auditory experience.

[0030] A speech segmentation unit is used to perform word segmentation processing on real-time user speech data; Obtain enhanced real-time user voice data; Based on natural language processing technology, the real-time user voice data is segmented; Cut the continuous real-time user voice data into a series of small semantic units that are easy to analyze; A feature extraction unit, used to extract features from real-time user voice data; Get real-time user voice data after word segmentation; Based on the Fourier transform method, the features of speech pronunciation, intonation, and speaking speed that are valuable for the early diagnosis of Parkinson's disease are extracted from the real-time user speech data after word segmentation, and the user speech feature data is determined.

[0031] Among them, the model training module is used to build an early diagnosis model for Parkinson's disease based on intelligent speech recognition analysis according to the early diagnosis needs of Parkinson's disease based on intelligent speech recognition analysis; In this embodiment, the model training module includes: A data collection unit, used to collect user voice history data; According to the needs of early diagnosis of Parkinson's disease based on intelligent speech recognition analysis, user voice history data is collected and divided; Determine the Parkinson's disease early diagnosis training set and Parkinson's disease early diagnosis test set; A model building unit, used to build an early diagnosis model for Parkinson's disease based on intelligent speech recognition analysis; According to the demand for early diagnosis of Parkinson's disease based on intelligent speech recognition analysis, a convolutional neural network model framework suitable for early diagnosis of Parkinson's disease is selected; Based on the Parkinson's disease early diagnosis training set, the selected convolutional neural network model framework suitable for the early diagnosis of Parkinson's disease is trained; An early diagnosis model for Parkinson's disease based on intelligent speech recognition analysis was determined.

[0032] Among them, the test optimization module is used to test and optimize the Parkinson's disease early diagnosis model based on intelligent speech recognition analysis to determine the optimal Parkinson's disease early diagnosis model based on intelligent speech recognition analysis; In this embodiment, the test optimization module includes: A test and evaluation unit, used to test and evaluate the Parkinson's disease early diagnosis model based on intelligent speech recognition analysis; Obtain an early diagnosis model for Parkinson's disease based on intelligent speech recognition analysis; Based on the Parkinson's disease early diagnosis test set, the performance test and evaluation of the Parkinson's disease early diagnosis model based on intelligent speech recognition analysis was conducted; Determine the performance test evaluation results based on the Parkinson's disease early diagnosis model; An adjustment and optimization unit, used to adjust and optimize the Parkinson's disease early diagnosis model based on intelligent speech recognition analysis; Obtain performance test evaluation results based on the Parkinson's disease early diagnosis model; According to the performance test evaluation results of the Parkinson's disease early diagnosis model, the Parkinson's disease early diagnosis model based on intelligent speech recognition analysis is mined and analyzed; Determine the parameter adjustment optimization scheme based on the early diagnosis model of Parkinson's disease; According to the parameter adjustment optimization scheme based on the Parkinson's disease early diagnosis model, the parameters of the Parkinson's disease early diagnosis model based on intelligent speech recognition analysis are adjusted and optimized; Determine the optimal early diagnosis model for Parkinson's disease based on intelligent speech recognition analysis.

[0033] Among them, the analysis and diagnosis module is used to mine and analyze the user's voice feature data based on the optimal Parkinson's disease early diagnosis model based on intelligent voice recognition analysis, perform early diagnosis of Parkinson's disease on the user, and determine the user's Parkinson's disease early diagnosis result based on intelligent voice recognition analysis; In this embodiment, the analysis and diagnosis module includes: A data extraction unit, used to extract user voice feature data; An analysis and diagnosis unit, used for early diagnosis of Parkinson's disease for users; Obtain the best early diagnosis model for Parkinson's disease based on intelligent speech recognition analysis; Input the user's voice feature data into the optimal Parkinson's disease early diagnosis model based on intelligent speech recognition analysis; Based on the optimal Parkinson's disease early diagnosis model based on intelligent speech recognition analysis, the user's speech feature data is mined and analyzed to perform early diagnosis of Parkinson's disease for the user; Determine the early diagnosis results of users' Parkinson's disease based on intelligent speech recognition analysis.

[0034] Among them, the display management module is used to warn users of early diagnosis of Parkinson's disease and to display and manage the results of early diagnosis of Parkinson's disease in users.

[0035] In this embodiment, the display management module includes: Diagnosis and early warning unit, used to warn users of early diagnosis of Parkinson's disease; Obtain early diagnosis results of users’ Parkinson’s disease based on intelligent speech recognition analysis; According to the early diagnosis results of Parkinson's disease of users based on intelligent speech recognition analysis, timely warning of early diagnosis of Parkinson's disease of users can be given, so that users can receive follow-up treatment in time; A display management unit, used to display and manage the user's Parkinson's disease early diagnosis results; Obtain early diagnosis results of users’ Parkinson’s disease based on intelligent speech recognition analysis; The early diagnosis results of the user's Parkinson's disease based on the intelligent speech recognition analysis are displayed to the user in a visual form, and the early diagnosis results of the user's Parkinson's disease based on the intelligent speech recognition analysis are stored and managed.

[0036] In summary, by collecting real-time user voice data and processing the real-time user voice data, the user voice feature data is determined, and according to the early diagnosis needs of Parkinson's disease based on intelligent voice recognition analysis, a Parkinson's disease early diagnosis model based on intelligent voice recognition analysis is constructed, and by testing and optimizing the Parkinson's disease early diagnosis model based on intelligent voice recognition analysis, the optimal Parkinson's disease early diagnosis model based on intelligent voice recognition analysis is determined, and based on the optimal Parkinson's disease early diagnosis model based on intelligent voice recognition analysis, the user voice feature data is mined and analyzed, and Parkinson's disease is early diagnosed for the user, and the early diagnosis results of Parkinson's disease for the user based on intelligent voice recognition analysis are determined, and early diagnosis of Parkinson's disease for the user is warned, and the early diagnosis results of Parkinson's disease for the user are displayed and managed. Parkinson's disease can be diagnosed early based on intelligent voice recognition analysis, and early detection and treatment of Parkinson's disease can be achieved, which can improve the effect of early diagnosis of Parkinson's disease.

[0037] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0038] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. The Parkinson's disease early diagnosis system based on intelligent speech recognition analysis is characterized by: include: The voice collection module is used to monitor the user's recording in real time and capture the user's voice signal, and collect the user's real-time voice data; The voice processing module is used to perform noise reduction, enhancement, word segmentation and feature extraction on the real-time user voice data to determine the user voice feature data; A model training module is used to build an early diagnosis model for Parkinson's disease based on intelligent speech recognition analysis according to the early diagnosis needs of Parkinson's disease based on intelligent speech recognition analysis; A test optimization module is used to test and optimize the Parkinson's disease early diagnosis model based on intelligent speech recognition analysis to determine the optimal Parkinson's disease early diagnosis model based on intelligent speech recognition analysis; An analysis and diagnosis module is used to mine and analyze the user's voice feature data based on the optimal Parkinson's disease early diagnosis model based on intelligent voice recognition analysis, to perform early diagnosis of Parkinson's disease on the user, and to determine the user's Parkinson's disease early diagnosis result based on intelligent voice recognition analysis; The display management module is used to warn users of early diagnosis of Parkinson's disease and to display and manage the results of early diagnosis of Parkinson's disease for users.

2. The Parkinson's disease early diagnosis system based on intelligent speech recognition analysis according to claim 1 is characterized in that: The voice collection module comprises: A voice capture unit, used to monitor user recordings in real time and capture user voice signals; According to the needs of early diagnosis of Parkinson's disease based on intelligent speech recognition analysis, real-time monitoring of user recordings is carried out to capture user voice signals; Voice collection unit, used to collect real-time data of user voice; Acquire user voice signals, collect user voice signals, and determine user voice real-time data.

3. The Parkinson's disease early diagnosis system based on intelligent speech recognition analysis according to claim 2 is characterized in that: The speech processing module comprises: A speech noise reduction unit is used to perform noise reduction processing on real-time user speech data; Get real-time user voice data; Based on the median filter, the real-time user voice data is de-noised; Remove noise from real-time user voice data; A speech enhancement unit is used to enhance the real-time data of user speech; Obtain real-time user voice data after noise reduction; The speech enhancement algorithm based on Kalman filtering can enhance the real-time user speech data after noise reduction, suppress and reduce noise interference; Real-time user speech data valuable for early diagnosis of Parkinson's disease is extracted from the noise background.

4. The Parkinson's disease early diagnosis system based on intelligent speech recognition analysis according to claim 3 is characterized in that: The speech processing module further includes: A first volume data extraction module is used to extract the volume of each frame of audio in the real-time voice data; A first comparison module, used for comparing the volume of each frame of audio with a preset volume threshold; An audio frame number extraction module is used to extract the number of audio frames whose volume is lower than a preset volume threshold; A second comparison module, used for comparing the number of audio frames with a preset frame number threshold; A second volume data extraction module is used to retrieve the volume value corresponding to the audio below the preset volume threshold when the number of audio frames exceeds the preset frame number threshold; A volume index value acquisition module, used to obtain a volume index value using the volume value corresponding to the audio that is lower than the preset volume threshold; The volume index value is obtained by the following formula: Wherein, K represents the volume index value; n represents the number of audio frames below the preset volume threshold; F i Indicates the volume value of the audio of the i-th frame that is lower than the preset volume threshold; F y Indicates the preset volume threshold; F max Indicates the maximum volume of the audio that is lower than the preset volume threshold; F min Indicates the minimum volume value of the audio that is lower than the preset volume threshold; F m represents the preset minimum allowed volume value; f represents the adjustment coefficient, and the adjustment coefficient is obtained by the following formula: Wherein, f represents the adjustment coefficient; n represents the number of audio frames below the preset volume threshold; F i Indicates the volume value of the audio of the i-th frame that is lower than the preset volume threshold; F y Indicates the preset volume threshold; F m Indicates the preset minimum allowed volume value; A third comparison module, used for comparing the volume index value with a preset index value threshold; The volume adjustment module is used to adjust the volume of the audio when the volume index value is lower than a preset index value threshold.

5. The Parkinson's disease early diagnosis system based on intelligent speech recognition analysis according to claim 4 is characterized in that: The volume adjustment module comprises: A first audio data acquisition module is used to extract the audio volume that exceeds the preset volume threshold and has the closest time distance before and after each frame of audio that is lower than the preset volume threshold as the first audio data; A second audio data acquisition module, used to extract the volume of each frame of audio below the preset volume threshold as second audio data; A volume index value extraction module, used to extract the volume index value; A volume enhancement adjustment module, configured to perform volume enhancement adjustment on each frame of audio that is lower than the preset volume threshold using the first audio data, the second audio data and the volume index value; The volume of each frame of audio that is lower than the preset volume threshold after enhancement and adjustment is obtained by the following formula: Wherein, F represents the volume of each frame of audio that is lower than the preset volume threshold after adjustment; F0 represents the volume of each frame of audio that is lower than the preset volume threshold; F x represents the volume of the audio frame that is closer to the preset volume threshold than the audio frame that is closer to the preset volume threshold; K represents the volume index value; f represents the adjustment coefficient; f represents the adjustment coefficient; y represents the adjustment factor, and the adjustment factor is obtained by the following formula: Among them, f y represents the adjustment factor; n represents the number of audio frames below the preset volume threshold; F i Indicates the volume value of the audio of the i-th frame that is lower than the preset volume threshold; F y Indicates the preset volume threshold.

6. The Parkinson's disease early diagnosis system based on intelligent speech recognition analysis according to claim 3 is characterized in that: The speech processing module further includes: A speech segmentation unit is used to perform word segmentation processing on real-time user speech data; Obtain enhanced real-time user voice data; Based on natural language processing technology, the real-time user voice data is segmented; Divide the continuous real-time user voice data into a series of semantic units that are easy to analyze; A feature extraction unit, used to extract features from real-time user voice data; Get real-time user voice data after word segmentation; Based on the Fourier transform method, the features of speech pronunciation, intonation, and speaking speed that are valuable for the early diagnosis of Parkinson's disease are extracted from the real-time user speech data after word segmentation, and the user speech feature data is determined.

7. The Parkinson's disease early diagnosis system based on intelligent speech recognition analysis according to claim 6 is characterized in that: The model training module includes: A data collection unit, used to collect user voice history data; According to the needs of early diagnosis of Parkinson's disease based on intelligent speech recognition analysis, user voice history data is collected and divided; Determine the Parkinson's disease early diagnosis training set and Parkinson's disease early diagnosis test set; A model building unit, used to build an early diagnosis model for Parkinson's disease based on intelligent speech recognition analysis; According to the demand for early diagnosis of Parkinson's disease based on intelligent speech recognition analysis, a convolutional neural network model framework suitable for early diagnosis of Parkinson's disease is selected; Based on the Parkinson's disease early diagnosis training set, the selected convolutional neural network model framework suitable for the early diagnosis of Parkinson's disease is trained; An early diagnosis model for Parkinson's disease based on intelligent speech recognition analysis was determined.

8. The Parkinson's disease early diagnosis system based on intelligent speech recognition analysis according to claim 7 is characterized in that: The test optimization module comprises: A test and evaluation unit, used to test and evaluate the Parkinson's disease early diagnosis model based on intelligent speech recognition analysis; Obtain an early diagnosis model for Parkinson's disease based on intelligent speech recognition analysis; Based on the Parkinson's disease early diagnosis test set, the performance test and evaluation of the Parkinson's disease early diagnosis model based on intelligent speech recognition analysis was conducted; Determine the performance test evaluation results based on the Parkinson's disease early diagnosis model; An adjustment and optimization unit, used to adjust and optimize the Parkinson's disease early diagnosis model based on intelligent speech recognition analysis; Obtain performance test evaluation results based on the Parkinson's disease early diagnosis model; According to the performance test evaluation results of the Parkinson's disease early diagnosis model, the Parkinson's disease early diagnosis model based on intelligent speech recognition analysis is mined and analyzed; Determine the parameter adjustment optimization scheme based on the early diagnosis model of Parkinson's disease; According to the parameter adjustment optimization scheme based on the Parkinson's disease early diagnosis model, the parameters of the Parkinson's disease early diagnosis model based on intelligent speech recognition analysis are adjusted and optimized; Determine the optimal early diagnosis model for Parkinson's disease based on intelligent speech recognition analysis.

9. The Parkinson's disease early diagnosis system based on intelligent speech recognition analysis according to claim 8, characterized in that: The analysis and diagnosis module comprises: A data extraction unit, used to extract user voice feature data; An analysis and diagnosis unit, used for early diagnosis of Parkinson's disease for users; Obtain the best early diagnosis model for Parkinson's disease based on intelligent speech recognition analysis; Input the user's voice feature data into the optimal Parkinson's disease early diagnosis model based on intelligent speech recognition analysis; Based on the optimal Parkinson's disease early diagnosis model based on intelligent speech recognition analysis, the user's speech feature data is mined and analyzed to perform early diagnosis of Parkinson's disease for the user; Determine the early diagnosis results of users' Parkinson's disease based on intelligent speech recognition analysis.

10. The Parkinson's disease early diagnosis system based on intelligent speech recognition analysis according to claim 9, characterized in that: The display management module includes: Diagnosis and early warning unit, used to warn users of early diagnosis of Parkinson's disease; Obtain early diagnosis results of users’ Parkinson’s disease based on intelligent speech recognition analysis; According to the early diagnosis results of Parkinson's disease of users based on intelligent speech recognition analysis, timely warning of early diagnosis of Parkinson's disease of users can be given, so that users can receive follow-up treatment in time; A display management unit, used to display and manage the user's Parkinson's disease early diagnosis results; Obtain early diagnosis results of users’ Parkinson’s disease based on intelligent speech recognition analysis; The early diagnosis results of the user's Parkinson's disease based on the intelligent speech recognition analysis are displayed to the user in a visual form, and the early diagnosis results of the user's Parkinson's disease based on the intelligent speech recognition analysis are stored and managed.

Images