Construction system and method of facial expression recognition training model

Abstract

The invention discloses a construction system and method for a facial expression recognition training model, and the system comprises a database assembly, the database assembly is composed of a data collection module, a data classification module, a data storage module, a data output module, a historical access module and a backup storage module, and the data collection module is in control connection with the data classification module; according to the invention, by using the database component, the data information in the backup storage module is directly called to be output, so that the data extraction efficiency during model training is improved, and the model training is accelerated; by using a feature extraction component, redundant features are removed, key features useful for expression classification are reserved, and feature extraction is optimized, so that the complexity of the features is reduced, and the robustness of subsequent model recognition output is guaranteed; by using the optimization output component, combining the prediction results of the plurality of models and performing weighting processing, the final recognition result is generated, and the stability and the accuracy of the model recognition result are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of facial expression recognition, and in particular to a system and method for constructing a facial expression recognition training model. Background Art

[0002] Facial recognition, also known as face recognition, facial image recognition or facial recognition, is a recognition method based on biometrics. It uses a general camera as a recognition information acquisition device to obtain the facial image of the recognition object in a non-contact manner, and compares the image with the database image after the computer system obtains the image to complete the recognition process. Facial expression recognition is a technology that identifies emotional state by analyzing facial expressions and movements. Facial expression recognition models are partially used for facial expression recognition, but traditional facial expression recognition models still have shortcomings. First, when training traditional expression recognition models, large amounts of data need to be frequently extracted, and the extraction efficiency is insufficient, which slows down model training; second, when training traditional expression recognition models, the extracted features are relatively complex and cannot guarantee the robustness of subsequent model recognition outputs; third, when training traditional expression recognition models, high requirements are placed on the stability and accuracy of model recognition results; therefore, it is necessary to design a system and method for building a facial expression recognition training model. Summary of the invention

[0003] The purpose of the present invention is to provide a system and method for constructing a facial expression recognition training model to solve the problems raised in the above background technology.

[0004] To achieve the above-mentioned purpose, the present invention provides the following technical scheme: a system for constructing a facial expression recognition training model, comprising a database component, wherein the database component is composed of a data acquisition module, a data classification module, a data storage module, a data output module, a history access module and a backup storage module, the data acquisition module is controlled to be connected to the data classification module, the data classification module is controlled to be connected to the data storage module, the data storage module is controlled to be connected to the data output module, the data output module is controlled to be connected to the history access module, the history access module is controlled to be connected to the backup storage module, and the data output module is controlled to be connected to the data extraction module in the data extraction component.

[0005] As a further technical solution of the present invention, the data extraction component is composed of a data extraction module, an expression recognition module, a data processing module, a standardization module, a normalization module, a grayscale processing module, a data enhancement module and a data docking module. The data extraction module controls the connection with the expression recognition module, the expression recognition module controls the connection with the data processing module, the data processing module controls the connection with the standardization module, the data processing module controls the connection with the normalization module, the data processing module controls the connection with the grayscale processing module, the data processing module controls the connection with the data enhancement module, and the data processing module controls the connection with the data docking module.

[0006] As a further technical solution of the present invention, the data docking module controls the data receiving module in the feature extraction component.

[0007] As a further technical solution of the present invention, the feature extraction component is composed of a data receiving module, an organ feature module, a texture area module, a predefined module, a feature recognition module, a geometric positioning module, a learning extraction module, a redundant processing module, a feature selection module, a dimensionality reduction technology module and a feature output module. The data receiving module is controlled to be connected with the organ feature module, the organ feature module is controlled to be connected with the feature recognition module, the feature recognition module is controlled to be connected with the texture area module, the texture area module is controlled to be connected with the data receiving module, the data receiving module is controlled to be connected with the predefined module, the predefined module is controlled to be connected with the feature recognition module, the feature recognition module is controlled to be connected with the geometric positioning module, the geometric positioning module is controlled to be connected with the redundant processing module, the redundant processing module is controlled to be connected with the learning extraction module, and the learning extraction module is controlled to be connected with the feature recognition module.

[0008] As a further technical solution of the present invention, the redundant processing module controls the connection with a feature selection module, the redundant processing module controls the connection with a dimensionality reduction technology module, the redundant processing module controls the connection with a feature output module, and the feature output module controls the connection with a feature receiving module in the optimization output component.

[0009] As a further technical solution of the present invention, the optimization output component is composed of a feature receiving module, a receiving and combining module, an integrated learning module, a weighted average module, a result comparison module, a model adjustment module and a model output module. The feature receiving module controls the connection with the receiving and combining module, the receiving and combining module controls the connection with the integrated learning module, the integrated learning module controls the connection with the weighted average module, the weighted average module controls the connection with the result comparison module, the result comparison module controls the connection with the model adjustment module, and the model adjustment module controls the connection with the model output module.

[0010] A method for constructing a facial expression recognition training model comprises the following steps: step 1: database establishment and data extraction processing; step 2: model feature extraction training; step 3: model optimization output;

[0011] In the above step 1, a database for model training is established and extraction processing is performed;

[0012] In the above step 2, the features in the data are extracted and trained;

[0013] In the above step three, the results of the model output are evaluated and the model is optimized and output.

[0014] As a further technical solution of the present invention, in the step one, various facial image data information is collected through the data acquisition module, and classified and defined through the data classification module, different expression types are defined, and transmitted to the data storage module for storage, completing the establishment of the model training database, and the data in the data storage module is transmitted according to the request through the data output module, and the frequency and regularity of data requests are recorded through the historical access module, and the high-frequency access data is stored through the backup storage module. When the same data request appears next time, the data information in the backup storage module is directly called for output; the data in the established database is extracted through the data extraction module, and transmitted to the expression recognition module for expression recognition, as a comparison reference for subsequent model output, the data is transmitted to the data processing module, and the data is pre-processed through the standardization module, the normalization module, the grayscale processing module and the data enhancement module to reduce noise and interference in the data information, and the processed data information is transmitted to the data docking module.

[0015] As a further technical solution of the present invention, in the step 2, the processed data is transmitted to the data receiving module through the data docking module, and the received data is feature located and extracted through the organ feature module, the texture area module and the predefined module to extract information that is beneficial to expression classification, and the processed data is transmitted to the feature recognition module, and the feature recognition module uses the geometric positioning module to locate the key feature points of the face, and then calculates the distance, angle and other geometric parameters between these feature points, which are used as expression features, and the hierarchical features in the image are automatically learned through the learning extraction module to learn the feature representation that is useful for expression classification, and the extracted and recognized expression feature data is transmitted to the redundant processing module, and the redundant processing module is used to perform principal component analysis PCA and linear discriminant analysis LDA through the feature selection module and the dimensionality reduction technology module to remove redundant features and retain the key features useful for expression classification, and the processed expression classification information is output through the feature output module.

[0016] As a further technical solution of the present invention, in the step three, the expression recognition information extracted and output by the model is received through the feature receiving module, the expression recognition information output by multiple models is combined through the receiving and combining module, and transmitted to the integrated learning module through the receiving and combining module, and the combined expression information is transmitted to the weighted average module through the integrated learning module, and the expression recognition information of each model is weighted in the weighted average module to form a final unique expression recognition information and output it to the result comparison module, which is compared with the classified information in the database and judged according to the set recognition error threshold. If it is unqualified, the model recognition parameters are adjusted through the model adjustment module, and the recognition training is re-performed, and it is output through the model output module after it is qualified.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: the system and method for constructing a facial expression recognition training model collect various face image data information through a data acquisition module, classify and define them through a data classification module, define different expression types, and transmit them to a data storage module for storage, complete the establishment of a model training database, and transmit the data in the data storage module according to the request through a data output module, and record the frequency and regularity of data requests through a historical access module, and store the data with high frequency access through a backup storage module, and when the same data request appears next time, directly call the data information in the backup storage module for output, thereby improving the efficiency of data extraction during model training and accelerating model training; the processed data is transmitted to a data receiving module through a data docking module, and the received data is feature located and extracted through an organ feature module, a texture area module and a predefined module, and information that is beneficial to expression classification is extracted, and the processed data is transmitted to a feature recognition module, and the key feature points of the face are located by using a geometric positioning module through the feature recognition module, and then the relationship between these feature points is calculated. The geometric parameters such as the distance and angle between the two images are used as expression features. The hierarchical features in the image are automatically learned through the learning and extraction module, and the feature representation useful for expression classification is learned. The extracted and recognized expression feature data is transmitted to the redundant processing module. The redundant processing module is subjected to principal component analysis PCA and linear discriminant analysis LDA through the feature selection module and the dimensionality reduction technology module to remove redundant features, retain key features useful for expression classification, optimize feature extraction, reduce feature complexity, and ensure the robustness of subsequent model recognition output; the expression recognition information output by the model extraction is received through the feature receiving module, the expression recognition information output by multiple models is combined through the receiving and combining module, and transmitted to the integrated learning module through the receiving and combining module, and the combined expression information is transmitted to the weighted average module through the integrated learning module, and the expression recognition information of each model is weighted in the weighted average module to form the final unique expression recognition information to be output to the result comparison module, and the prediction results of multiple models are combined and weighted to generate the final recognition result, thereby improving the stability and accuracy of the model recognition result. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is the overall architecture diagram of the present invention;

[0019] Figure 2 is a flow chart of the database component in the present invention;

[0020] Figure 3 is a flow chart of the data extraction component in the present invention;

[0021] Figure 4is a flow chart of the feature extraction component in the present invention;

[0022] Figure 5 A flow chart of optimizing the output component in the present invention;

[0023] Figure 6 is a flow chart of the method of the present invention;

[0024] In the figure: 1. Database component; 2. Data extraction component; 3. Feature extraction component; 5. Optimization output component; 11. Data acquisition module; 12. Data classification module; 13. Data storage module; 14. Data output module; 15. History access module; 16. Backup storage module; 21. Data extraction module; 22. Expression recognition module; 23. Data processing module; 24. Standardization module; 25. Normalization module; 26. Grayscale processing module; 27. Data enhancement module; 28. Data output module; 3 1. Data receiving module; 32. Organ feature module; 33. Texture area module; 34. Predefined module; 35. Feature recognition module; 36. Geometric positioning module; 37. Learning extraction module; 38. Redundancy processing module; 39. Feature selection module; 40. Dimensionality reduction technology module; 41. Feature output module; 51. Feature receiving module; 52. Receiving and combining module; 53. Integrated learning module; 54. Weighted average module; 55. Result comparison module; 56. Model adjustment module; 57. Model output module. DETAILED DESCRIPTION

[0025] 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.

[0026] Please see attached Figure 1 -Attached Figure 5An embodiment of the present invention provides a system for constructing a facial expression recognition training model, comprising a database component 1, the database component 1 is composed of a data acquisition module 11, a data classification module 12, a data storage module 13, a data output module 14, a history access module 15 and a backup storage module 16, the data acquisition module 11 is controlled to be connected to the data classification module 12, the data classification module 12 is controlled to be connected to the data storage module 13, the data storage module 13 is controlled to be connected to the data output module 14, the data output module 14 is controlled to be connected to the history access module 15, the history access module 15 is controlled to be connected to the backup storage module 16, the data output module 14 is controlled to be connected to the data extraction component 2 The data extraction module 21 in the data extraction component 2 is composed of a data extraction module 21, an expression recognition module 22, a data processing module 23, a standardization module 24, a normalization module 25, a grayscale processing module 26, a data enhancement module 27 and a data docking module 28. The data extraction module 21 is controlled to be connected with the expression recognition module 22, the expression recognition module 22 is controlled to be connected with the data processing module 23, the data processing module 23 is controlled to be connected with the standardization module 24, the data processing module 23 is controlled to be connected with the normalization module 25, the data processing module 23 is controlled to be connected with the grayscale processing module 26, the data processing module 23 is controlled to be connected with the data enhancement module 27, and the data processing module 23 is controlled to be connected with The data docking module 28 controls the connection with the data receiving module 31 in the feature extraction component 3. The feature extraction component 3 is composed of the data receiving module 31, the organ feature module 32, the texture region module 33, the predefined module 34, the feature recognition module 35, the geometric positioning module 36, the learning extraction module 37, the redundant processing module 38, the feature selection module 39, the dimension reduction technology module 40 and the feature output module 41. The data receiving module 31 controls the connection with the organ feature module 32, the organ feature module 32 controls the connection with the feature recognition module 35, the feature recognition module 35 controls the connection with the texture region module 33, and the texture region module 33 controls the connection with the data receiving module On block 31, the data receiving module 31 controls the connection with the predefined module 34, the predefined module 34 controls the connection with the feature recognition module 35, the feature recognition module 35 controls the connection with the geometric positioning module 36, the geometric positioning module 36 controls the connection with the redundant processing module 38, the redundant processing module 38 controls the connection with the learning extraction module 37, the learning extraction module 37 controls the connection with the feature recognition module 35, the redundant processing module 38 controls the connection with the feature selection module 39, the redundant processing module 38 controls the connection with the dimensionality reduction technology module 40, the redundant processing module 38 controls the connection with the feature output module 41, the feature output module 41 controls the connection with the feature receiving module 51 in the optimization output component 5,The optimization output component 5 is composed of a feature receiving module 51, a receiving and combining module 52, an integrated learning module 53, a weighted average module 54, a result comparison module 55, a model adjustment module 56 and a model output module 57. The feature receiving module 51 is controlled to be connected to the receiving and combining module 52, the receiving and combining module 52 is controlled to be connected to the integrated learning module 53, the integrated learning module 53 is controlled to be connected to the weighted average module 54, the weighted average module 54 is controlled to be connected to the result comparison module 55, the result comparison module 55 is controlled to be connected to the model adjustment module 56, and the model adjustment module 56 is controlled to be connected to the model output module 57.

[0027] Please see attached Figure 6 , an embodiment provided by the present invention: a method for constructing a facial expression recognition training model, comprising the following steps: step 1: database establishment and data extraction processing; step 2: model feature extraction training; step 3: model optimization output;

[0028] In the above step 1, a database for model training is established and extracted. Various facial image data information is collected through the data acquisition module 11, and classified and defined through the data classification module 12, different expression types are defined, and transmitted to the data storage module 13 for storage, completing the establishment of the model training database, and the data in the data storage module 13 is transmitted according to the request through the data output module 14, and the frequency and regularity of data requests are recorded through the historical access module 15, and the high-frequency access data is stored through the backup storage module 16. When the same data request appears next time, the data information in the backup storage module 16 is directly called for output; the data in the established database is extracted through the data extraction module 21, and transmitted to the expression recognition module 22 for expression recognition, as a comparison reference for subsequent model output, the data is transmitted to the data processing module 23, and the data is pre-processed through the standardization module 24, the normalization module 25, the grayscale processing module 26 and the data enhancement module 27 to reduce the noise and interference in the data information, and the processed data information is transmitted to the data docking module 28;

[0029] In the above step 2, the features in the data are extracted and trained, and the processed data is transmitted to the data receiving module 31 through the data docking module 28. The received data is located and extracted by the organ feature module 32, the texture area module 33 and the predefined module 34, and the information that is conducive to expression classification is extracted, and the processed data is transmitted to the feature recognition module 35. The feature recognition module 35 uses the geometric positioning module 36 to locate the key feature points of the face, and then calculates the distance, angle and other geometric parameters between these feature points, which are used as expression features. The hierarchical features in the image are automatically learned by the learning extraction module 37, and the feature representation useful for expression classification is learned, and the extracted and recognized expression feature data is transmitted to the redundancy processing module 38. The redundancy processing module 38 performs principal component analysis PCA and linear discriminant analysis LDA on it through the feature selection module 39 and the dimensionality reduction technology module 40, removes redundant features, retains the key features useful for expression classification, and outputs the processed expression classification information through the feature output module 41;

[0030] In the above step three, the results of the model output are evaluated and the model is optimized and output. The expression recognition information extracted and output by the model is received by the feature receiving module 51, and the expression recognition information output by multiple models is combined by the receiving and combining module 52, and transmitted to the integrated learning module 53 by the receiving and combining module 52. The combined expression information is transmitted to the weighted average module 54 by the integrated learning module 53. The expression recognition information of each model is weighted in the weighted average module 54 to form a final unique expression recognition information, which is output to the result comparison module 55, which is compared with the classified information in the database and judged according to the set recognition error threshold. If it is unqualified, the model recognition parameters are adjusted by the model adjustment module 56, and the recognition training is re-performed. After it is qualified, it is output through the model output module 57.

[0031] Working principle: When in use, various facial image data information is collected through the data acquisition module 11, and classified and defined through the data classification module 12, different expression types are defined, and transmitted to the data storage module 13 for storage, completing the establishment of the model training database, and the data in the data storage module 13 is transmitted according to the request through the data output module 14, and the frequency and regularity of data requests are recorded through the historical access module 15, and the high-frequency access data is stored through the backup storage module 16. When the same data request appears next time, the data information in the backup storage module 16 is directly called for output, which improves the efficiency of data extraction during model training and speeds up model training; the processed data is transmitted to the data receiving module 31 through the data docking module 28, and the received data is feature located and extracted through the organ feature module 32, the texture area module 33 and the predefined module 34, and the information that is conducive to expression classification is extracted, and the processed data is transmitted to the feature recognition module 35, and the key feature points of the face are located by the geometric positioning module 36 through the feature recognition module 35, and then the distance, angle and other geometric parameters between these feature points are calculated. Using this as the expression feature, the hierarchical features in the image are automatically learned through the learning and extraction module 37, and the feature representation useful for expression classification is learned, and the extracted and recognized expression feature data is transmitted to the redundant processing module 38. The redundant processing module 38 passes it through the feature selection module 39 and the dimensionality reduction technology module 40 to perform principal component analysis PCA and linear discriminant analysis LDA, remove redundant features, retain key features useful for expression classification, optimize feature extraction, reduce feature complexity, and ensure the robustness of subsequent model recognition output; receive the model feature through the feature receiving module 51 The output expression recognition information is taken, and the expression recognition information output by multiple models is combined through the receiving and combining module 52, and it is transmitted to the integrated learning module 53 through the receiving and combining module 52, and the combined expression information is transmitted to the weighted average module 54 through the integrated learning module 53. The expression recognition information of each model is weighted in the weighted average module 54 to form the final unique expression recognition information, which is output to the result comparison module 55. The prediction results of multiple models are combined and weighted processing is performed to generate the final recognition result, thereby improving the stability and accuracy of the model recognition results.

[0032] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A system for constructing a facial expression recognition training model, comprising a database component (1), characterized in that: The database component (1) is composed of a data acquisition module (11), a data classification module (12), a data storage module (13), a data output module (14), a history access module (15) and a backup storage module (16); the data acquisition module (11) is controlled to be connected to the data classification module (12); the data classification module (12) is controlled to be connected to the data storage module (13); the data storage module (13) is controlled to be connected to the data output module (14); the data output module (14) is controlled to be connected to the history access module (15); the history access module (15) is controlled to be connected to the backup storage module (16); and the data output module (14) is controlled to be connected to the data extraction module (21) in the data extraction component (2).

2. A system for constructing a facial expression recognition training model according to claim 1, characterized in that: The data extraction component (2) is composed of a data extraction module (21), an expression recognition module (22), a data processing module (23), a standardization module (24), a normalization module (25), a grayscale processing module (26), a data enhancement module (27) and a data docking module (28); the data extraction module (21) is controlled to be connected to the expression recognition module (22); the expression recognition module (22) is controlled to be connected to the data processing module (23); the data processing module (23) is controlled to be connected to the standardization module (24); the data processing module (23) is controlled to be connected to the normalization module (25); the data processing module (23) is controlled to be connected to the grayscale processing module (26); the data processing module (23) is controlled to be connected to the data enhancement module (27); and the data processing module (23) is controlled to be connected to the data docking module (28).

3. A system for constructing a facial expression recognition training model according to claim 2, characterized in that: The data docking module (28) controls the connection with the data receiving module (31) in the feature extraction component (3).

4. A system for constructing a facial expression recognition training model according to claim 3, characterized in that: The feature extraction component (3) is composed of a data receiving module (31), an organ feature module (32), a texture region module (33), a predefined module (34), a feature recognition module (35), a geometric positioning module (36), a learning extraction module (37), a redundant processing module (38), a feature selection module (39), a dimensionality reduction technology module (40) and a feature output module (41). The data receiving module (31) controls the connection with the organ feature module (32), the organ feature module (32) controls the connection with the feature recognition module (35), and the feature recognition module (35) controls The device is connected to a texture region module (33), the texture region module (33) controls the connection to a data receiving module (31), the data receiving module (31) controls the connection to a predefined module (34), the predefined module (34) controls the connection to a feature recognition module (35), the feature recognition module (35) controls the connection to a geometric positioning module (36), the geometric positioning module (36) controls the connection to a redundant processing module (38), the redundant processing module (38) controls the connection to a learning extraction module (37), and the learning extraction module (37) controls the connection to the feature recognition module (35).

5. A system for constructing a facial expression recognition training model according to claim 4, characterized in that: The redundant processing module (38) controls the connection with a feature selection module (39), the redundant processing module (38) controls the connection with a dimensionality reduction technology module (40), the redundant processing module (38) controls the connection with a feature output module (41), and the feature output module (41) controls the connection with a feature receiving module (51) in the optimization output component (5).

6. A system for constructing a facial expression recognition training model according to claim 5, characterized in that: The optimization output component (5) is composed of a feature receiving module (51), a receiving and combining module (52), an integrated learning module (53), a weighted average module (54), a result comparison module (55), a model adjustment module (56) and a model output module (57). The feature receiving module (51) is controlled to be connected to the receiving and combining module (52), the receiving and combining module (52) is controlled to be connected to the integrated learning module (53), the integrated learning module (53) is controlled to be connected to the weighted average module (54), the weighted average module (54) is controlled to be connected to the result comparison module (55), the result comparison module (55) is controlled to be connected to the model adjustment module (56), and the model adjustment module (56) is controlled to be connected to the model output module (57).

7. A method for constructing a facial expression recognition training model, comprising the following steps: Step 1: Database establishment and data extraction processing; Step 2: Model feature extraction training; Step 3: Model optimization output; It is characterized by: In the above step 1, a database for model training is established and extraction processing is performed; In the above step 2, the features in the data are extracted and trained; In the above step three, the results of the model output are evaluated and the model is optimized and output.

8. A method for constructing a facial expression recognition training model according to claim 7, characterized in that: In the step 1, various facial image data information is collected through the data collection module (11), and classified and defined through the data classification module (12), and different expression types are defined, and transmitted to the data storage module (13) for storage, so as to complete the establishment of the model training database, and the data in the data storage module (13) is transmitted according to the request through the data output module (14), and the frequency and regularity of data requests are recorded through the historical access module (15), and the data with high frequency access is stored through the backup storage module (16) until the next time the same data request occurs. When the user inputs the data, the data in the backup storage module (16) is directly called for output; the data in the established database is extracted by the data extraction module (21), and transmitted to the expression recognition module (22) for expression recognition, and the data is transmitted to the data processing module (23) as a comparison reference for the subsequent model output; the data is pre-processed by the standardization module (24), the normalization module (25), the grayscale processing module (26) and the data enhancement module (27) to reduce noise and interference in the data; and the processed data is transmitted to the data docking module (28).

9. The method for constructing a facial expression recognition training model according to claim 7, characterized in that: In the step 2, the processed data is transmitted to the data receiving module (31) through the data docking module (28), and the received data is characterized and extracted through the organ feature module (32), the texture area module (33) and the predefined module (34), and information that is conducive to expression classification is extracted, and the processed data is transmitted to the feature recognition module (35), and the feature recognition module (35) uses the geometric positioning module (36) to locate the key feature points of the face, and then calculates the geometric parameters such as the distance and angle between these feature points, which are used as expression features, and the hierarchical features in the image are automatically learned through the learning extraction module (37) to learn the feature representation useful for expression classification, and the extracted and recognized expression feature data is transmitted to the redundant processing module (38), and the redundant processing module (38) is subjected to principal component analysis PCA and linear discriminant analysis LDA through the feature selection module (39) and the dimensionality reduction technology module (40), and redundant features are removed, and key features useful for expression classification are retained, and the processed expression classification information is output through the feature output module (41).

10. The method for constructing a facial expression recognition training model according to claim 7, characterized in that: In the step three, the expression recognition information extracted and output by the model is received by the feature receiving module (51), the expression recognition information output by multiple models is combined by the receiving and combining module (52), the expression recognition information is transmitted to the integrated learning module (53) by the receiving and combining module (52), and the combined expression information is transmitted to the weighted average module (54) by the integrated learning module (53), the expression recognition information of each model is weighted in the weighted average module (54), and the final unique expression recognition information is output to the result comparison module (55), which is compared with the classified information in the database, and judged according to the set recognition error threshold. If it is unqualified, the model recognition parameters are adjusted by the model adjustment module (56), and the recognition training is re-performed. After it is qualified, it is output through the model output module (57).

Images