A face recognition system based on digital image processing technology

Abstract

The present invention provides a face recognition system based on digital image processing technology, including: a control area module: segmenting the target range through a preset attribute threshold to determine the corresponding control area; a face data acquisition module: collecting face images through an image sensor preset in the control area, and establishing a face image sample library; an intelligent environmental control module: detecting the face image sample library, extracting face point coordinates through feature training, and determining the same face image set; a comparison server module: optimally outputting the same face image set, processing it using a comparison server, generating a similarity score, and transmitting it to a preset security identification center; a security identification module: used to collect the similarity score, judge the identification result through a preset threshold, and form a corresponding management plan. The present invention uses digital image processing technology to achieve a high-precision face recognition process and effectively ensure the security of the target range.

Description

Technical Field

[0001] The present invention relates to the fields of digital image processing technology and artificial intelligence technology, and in particular to a face recognition system based on digital image processing technology. Background Art

[0002] In daily life, banks, offices, confidential bases and schools require strict personnel verification procedures. However, these places currently rely on traditional methods such as manual inspection and document verification for verification. Traditional verification methods rely on human eye comparison and memory search, which can only make simple judgments but not accurate identification. At the same time, they cannot respond in time when encountering emergencies. It not only wastes a lot of manpower and material resources, but also takes a long time to verify. In addition, it is inefficient and inaccurate, which can easily lead to misjudgment. Banks, offices, confidential bases and schools need a new type of intelligent face recognition system that focuses on the target object. Summary of the invention

[0003] The present invention provides a face recognition system based on digital image processing technology, which is used to solve the problems in the above-mentioned background technology.

[0004] The present invention provides a face recognition system based on digital image processing technology, comprising:

[0005] Control area module: divides the control area by a preset attribute threshold value to determine the corresponding target range; wherein the attribute threshold value represents the attribute threshold value of the display attribute of the control area;

[0006] Face data acquisition module: collects face images through image sensors preset within the target range, and establishes a face image sample library based on the face images;

[0007] Intelligent environment control module: detects the face image sample library, extracts the face point coordinates of the face image to be identified through feature training, and determines the same face image set;

[0008] Comparison server module: optimally output the same face image set, process it using the comparison server, generate a similarity score, and transmit the similarity score to a preset security identification center;

[0009] Security identification module: used to collect the similarity scores, judge the identification results through preset thresholds, and form a corresponding management plan.

[0010] As an embodiment of the present technical solution, the control area module includes:

[0011] Area setting unit: used to set up an integrated control system in the preset terminal equipment, form an integrated range in the control area, and build a data sharing platform;

[0012] Parameter acquisition unit: used to acquire environmental parameters within the integration range, upload them to the data sharing platform, identify their display attributes, and generate a parameter catalog; wherein,

[0013] The display attributes include at least texture attributes, object attributes and detail attributes within the integration range;

[0014] Control area unit: used to adapt the parameter catalog according to the preset attribute threshold value and determine the corresponding target range through segmentation; wherein,

[0015] The control area at least includes a cross control area and a separation control area.

[0016] As an embodiment of the present technical solution, the face data acquisition module includes:

[0017] A face acquisition unit is used to acquire a real-time image containing face features in the control area through an image sensor preset in the control area; wherein:

[0018] The image sensor includes at least one or more of a face recognition camera, a face recognition camera, a face recognition gate and a PDA data collector;

[0019] Image pre-screening unit: used to pre-screen the real-time image based on image processing technology, remove images with insufficient facial features in the real-time image, and determine the facial image to be processed; wherein,

[0020] The preprocessing technology at least includes geometric correction, image filtering, image enhancement and image edge detection methods;

[0021] Image sample library unit: used for transmitting the face image to be processed to the data sharing platform and integrating it into a face image sample library.

[0022] As an embodiment of the present technical solution, the intelligent environmental control module includes:

[0023] A face detection unit is used to input the face image sample library into a preset detection network to obtain the position of the face candidate frame in the face image to be processed;

[0024] Feature training unit: used for performing convolutional neural network learning on the position of the face candidate frame to determine the coordinates of the face feature points;

[0025] The same face unit is used to process the face image to be processed based on the coordinates of the face feature points by applying a classification method to form a set of the same face images; wherein,

[0026] The classification method at least includes a maximum likelihood classification method and a minimum distance classification method;

[0027] The feature training unit comprises:

[0028] The first convolutional network subunit is used for capturing the face candidate frame image unit based on the position of the face candidate frame to generate a candidate frame screenshot;

[0029] The second convolutional network subunit is used to integrate the candidate frame screenshots, build a face recognition network based on the face image sample library, and adjust the parameters of the face recognition network to build a face recognition model;

[0030] Feature point extraction subunit: input the face image to be processed into the face recognition model to obtain the coordinates of the face feature points.

[0031] As an embodiment of the present technical solution, the comparison server module includes:

[0032] Feature point acquisition unit: used to acquire all the facial images to be processed of the same face and extract the coordinates of standard facial feature points;

[0033] A calculation unit is used to obtain pixel points of the face image to be processed, project each pixel point on the face image to be processed into a preset parameter domain, and obtain the filtered pixel points;

[0034]

[0035] in Represents the first face image to be processed after screening Line Column pixels; Representative Line The pixel value of the pixel point in the column, Represents the preset parameter domain for face images, represents the inverse of the preset parameter domain for face images, Representative Line The pixel value after the pixel point of the column is transposed, Represents the preset parameter domain for face images The ideal range of values ​​for the rows, Represents the preset parameter domain for face images The ideal range of values ​​for the column;

[0036] Accumulated peak unit: used to accumulate the pixel points of the face image to be processed through a preset sampling period and calculate the accumulated peak value;

[0037]

[0038] in, represents the cumulative peak value, represent No. The sampling period of the pixels in a row, represent No. The sampling period of the pixel points of the column;

[0039] Standard facial feature point coordinate unit: used to analyze the points corresponding to the accumulated peak values ​​and map them to the facial image parameter space to obtain the standard facial feature point coordinates;

[0040] Constant matrix unit: used to constantize the coordinates of the standard facial feature points and calculate the constant matrix of the coordinates of the standard facial feature points;

[0041] A quality numerical result unit is used to estimate the parameters of the constant matrix of all the facial images to be processed of the same face by using a preset quality evaluation index, and determine the quality numerical results of all the facial images to be processed of the same face;

[0042] A quality evaluation unit: used for sorting the quality value results, determining the sorting result, and outputting the best face image based on the sorting result;

[0043] Feature code unit: used for converting the best face image into a feature code in the central server, and transmitting the feature code to a preset comparison server;

[0044] Similarity calculation unit: used to calculate and compare the feature code with the pre-recorded face model in the comparison server, determine the similarity score, and transmit the similarity score to the security identification center.

[0045] As an embodiment of the present technical solution, the facial feature point coordinates include at least left eye coordinates, right eye coordinates, nose coordinates, left mouth corner coordinates and right mouth corner coordinates.

[0046] As an embodiment of the present technical solution, the optimal face image is an image with a frontal posture and covering the coordinates of all facial feature points.

[0047] As an embodiment of the present technical solution, the security identification module includes:

[0048] Collection unit: used to receive the similarity score, display the data and generate a report;

[0049] The judgment and recognition unit is used to compare and judge the report with a preset threshold, determine whether the optimal face image has a security anomaly, and obtain a recognition result; wherein,

[0050] When the similarity score is less than a preset threshold, the recognition result is recognition failure;

[0051] When the similarity score is greater than a preset threshold, the recognition result is a successful recognition;

[0052] Result management unit: used to respond to access control actions based on the identification results and implement relevant management solutions.

[0053] As an embodiment of the present technical solution, the result management unit includes:

[0054] Successful identification management subunit: when the identification result is successful, the permission is granted, the corresponding door is controlled to open, and the face image is reversely propagated to restore it to the same face image set, and the capacity of the face image sample library is supplemented;

[0055] Identification failure management subunit: When the recognition result is recognition failure, no permission is granted, the corresponding access control is controlled to implement the prohibition action, and the best face image is reversely propagated to restore it to the same face image set, and the real-time position is tracked and locked and transmitted to the security terminal.

[0056] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0057] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0059] Figure 1 This is a module flow chart of a face recognition system based on digital image processing technology in an embodiment of the present invention;

[0060] Figure 2 This is a module flow chart of a face recognition system based on digital image processing technology in an embodiment of the present invention;

[0061] Figure 3 The module flow chart of a face recognition system based on digital image processing technology in an embodiment of the present invention. DETAILED DESCRIPTION

[0062] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0063] Embodiment 1:

[0064] like Figure 1 As shown, the present technical solution provides a face recognition system based on digital image processing technology, including:

[0065] Control area module: divides the control area by a preset attribute threshold value to determine the corresponding target range; wherein the attribute threshold value represents the attribute threshold value of the display attribute of the control area;

[0066] Face data acquisition module: collects face images through image sensors preset within the target range, and establishes a face image sample library based on the face images;

[0067] Intelligent environment control module: detects the face image sample library, extracts the face point coordinates of the face image to be identified through feature training, and determines the same face image set;

[0068] Comparison server module: optimally output the same face image set, process it using the comparison server, generate a similarity score, and transmit the similarity score to a preset security identification center;

[0069] Security identification module: used to collect the similarity scores, judge the identification results through preset thresholds, and form a corresponding management plan.

[0070] The working principle and beneficial effects of the above technical solution are:

[0071] In this technical solution, the control area module divides the target range, including offices, confidential bases and school entrances, through preset attribute thresholds, determines the corresponding control areas, and ensures the effectiveness of image acquisition through separate control and cross control; the face data acquisition module collects face images through image sensors preset in the control area, reflects the dynamic changes of the control area through real-time monitoring, and establishes a face image sample library based on face images to provide a reference for the next target object recognition and enhance the system's responsiveness; the intelligent environmental control module detects the face image sample library, extracts the face point coordinates of the face image to be recognized through feature training, determines the same face image set, and optimizes the recognition process by locking the key parameter coordinates. sequence, shortening the reaction time, and integrating all the collected images to reduce errors; the comparison server module optimally outputs the same face image set, improves the effective input of the target object face in the target range, and uses the comparison server for processing to generate a similarity score, and transmits the similarity score to the preset security recognition center. After digital image processing of the real-time face image and the pre-entered face model, the reliability of recognition is guaranteed; the security recognition module is used to collect similarity scores, judge the recognition results through a preset threshold, and form a corresponding management plan. The system strengthens the recognition control of the target object through differentiated responses to recognition success and recognition failure, and timely avoids risks to promote the construction of site safety. As a specific embodiment of the present technical solution, for example, at the face clocking-in place in the office, it is possible to determine whether the office is staff by recognizing the face, thereby ensuring the safety of the office area.

[0072] Embodiment 2:

[0073] According to the attached Figure 2 As shown, in one embodiment, the control area module includes:

[0074] Area setting unit: used to set up an integrated control system in the preset terminal equipment, form an integrated range in the control area, and build a data sharing platform;

[0075] Parameter acquisition unit: used to acquire environmental parameters within the integration range, upload them to the data sharing platform, identify their display attributes, and generate a parameter catalog; wherein,

[0076] The display attributes include at least texture attributes, object attributes and detail attributes within the integration range;

[0077] Control area unit: used to adapt the parameter catalog according to the preset attribute threshold value and determine the corresponding target range through segmentation; wherein,

[0078] The control area at least includes a cross control area and a separation control area.

[0079] The working principle and beneficial effects of the above technical solution are:

[0080] In this technical solution, the area setting unit is used to set up an integrated control system in a preset terminal device, forming an integrated control area in the target range, including offices, confidential bases and school entrances, and building a data sharing platform. By building an integrated control area and a data sharing platform, the site environment is intelligently coordinated to reduce the omission of site environment information; the parameter acquisition unit is used to obtain the environmental parameters of the integrated control area and upload it to the data sharing platform, and generate a parameter directory by marking its display attributes, including textures, objects and details, for easy background management; the control area unit is used to adapt the parameter directory according to the preset attribute threshold, summarize it, and determine the control area including the cross control area and the separated control area through segmentation to ensure the coverage of the target range. As an embodiment of the present technical solution, this face recognition system can also be applied to confidential bases, set at the door card, improve the confidentiality of the base, and improve the monitoring security of confidential bases by collecting face data.

[0081] Embodiment 3:

[0082] According to the attached Figure 3 As shown, in one embodiment, the face data acquisition module includes:

[0083] A face acquisition unit is used to acquire a real-time image containing face features in the control area through an image sensor preset in the control area; wherein:

[0084] The image sensor includes at least one or more of a face recognition camera, a face recognition camera, a face recognition gate and a PDA data collector;

[0085] Image pre-screening unit: used to pre-screen the real-time image based on image processing technology, remove images with insufficient facial features in the real-time image, and determine the facial image to be processed; wherein,

[0086] The preprocessing technology at least includes geometric correction, image filtering, image enhancement and image edge detection methods;

[0087] Image sample library unit: used for transmitting the face image to be processed to the data sharing platform and integrating it into a face image sample library.

[0088] The working principle and beneficial effects of the above technical solution are:

[0089] In this technical solution, the face acquisition unit is used to obtain real-time images containing facial features in the control area through image sensors pre-installed in the control area, including one or more of face recognition cameras, face recognition cameras, face recognition gates and PDA data collectors, to ensure the timeliness of information transmission; the image pre-screening unit is used to pre-screen the real-time image based on image processing technologies including geometric correction, image filtering, image enhancement and image edge detection methods, eliminate images with insufficient facial features in the real-time image, reduce unnecessary storage operations in the system space, determine the face images to be processed, and improve the performance of the face recognition system; the image sample library unit: is used to transmit the face images to be processed to the data sharing platform and integrate them into the face image sample library. As a specific embodiment of the present technical solution, the present technical solution can also be used at the entrances of schools or communities to protect the safety management of personnel.

[0090] Embodiment 4:

[0091] In one embodiment, the intelligent environment control module includes:

[0092] A face detection unit is used to input the face image sample library into a preset detection network to obtain the position of the face candidate frame in the face image to be processed;

[0093] Feature training unit: used for performing convolutional neural network learning on the position of the face candidate frame to determine the coordinates of the face feature points;

[0094] The same face unit is used to process the face image to be processed based on the coordinates of the face feature points by applying a classification method to form a set of the same face images; wherein,

[0095] The classification method at least includes a maximum likelihood classification method and a minimum distance classification method;

[0096] The feature training unit comprises:

[0097] The first convolutional network subunit is used for capturing the face candidate frame image unit based on the position of the face candidate frame to generate a candidate frame screenshot;

[0098] The second convolutional network subunit is used to integrate the candidate frame screenshots, build a face recognition network based on the face image sample library, and adjust the parameters of the face recognition network to build a face recognition model;

[0099] Feature point extraction subunit: input the face image to be processed into the face recognition model to obtain the coordinates of the face feature points.

[0100] The working principle and beneficial effects of the above technical solution are:

[0101] The face detection unit is used to input the face image sample library into a preset detection network to obtain the position of the face candidate frame in the face image to be processed; the feature training unit is used to perform convolutional neural network learning on the position of the face candidate frame to determine the coordinates of the face feature points; the face feature point coordinates include at least the left eye coordinates, the right eye coordinates, the nose coordinates, the left corner of the mouth coordinates and the right corner of the mouth coordinates; the same face unit is used to process the face image to be processed based on the face feature point coordinates, apply a classification method to identify the surveillance video by frame, determine the frame image, realize refined security monitoring in units of frames, effectively lock abnormal conditions, and form a set of the same face images; the classification method includes at least the maximum likelihood classification method and the minimum distance classification method. In the present technical solution, the first convolutional network subunit is used to intercept the face candidate frame image unit based on the position of the face candidate frame, generate a candidate frame screenshot, receive and store the frame image, and form an image database; the second convolutional network subunit is used to integrate the candidate frame screenshots, build a face recognition network based on the face image sample library, and adjust the parameters of the face recognition network to construct a face recognition model to achieve refined security monitoring. The feature point extraction subunit inputs the face image to be processed into the face recognition model to obtain the coordinates of the face feature points and improve the accuracy of the data.

[0102] Embodiment 5:

[0103] In one embodiment, the comparison server module includes:

[0104] Feature point acquisition unit: used to acquire all the facial images to be processed of the same face and extract the coordinates of standard facial feature points;

[0105] A calculation unit is used to obtain pixel points of the face image to be processed, project each pixel point on the face image to be processed into a preset parameter domain, and obtain the filtered pixel points;

[0106]

[0107] The first one represents the face image to be processed after screening. Line Column pixels; Representative Line The pixel value of the pixel point in the column, Represents the preset parameter domain for face images, represents the inverse of the preset parameter domain for face images, Representative Line The pixel value after the pixel point of the column is transposed, Represents the preset parameter domain for face images The ideal range of values ​​for the rows, Represents the preset parameter domain for face images The ideal range of values ​​for the column;

[0108] Accumulated peak unit: used to accumulate the pixel points of the face image to be processed through a preset sampling period and calculate the accumulated peak value;

[0109]

[0110] in, represents the cumulative peak value, represent No. The sampling period of the pixels in a row, represent No. The sampling period of the pixel points of the column;

[0111] Standard facial feature point coordinate unit: used to analyze the points corresponding to the accumulated peak values ​​and map them to the facial image parameter space to obtain the standard facial feature point coordinates;

[0112] Constant matrix unit: used to constantize the coordinates of the standard facial feature points and calculate the constant matrix of the coordinates of the standard facial feature points;

[0113] A quality numerical result unit is used to estimate the parameters of the constant matrix of all the facial images to be processed of the same face by using a preset quality evaluation index, and determine the quality numerical results of all the facial images to be processed of the same face;

[0114] A quality evaluation unit: used for sorting the quality value results, determining the sorting result, and outputting the best face image based on the sorting result;

[0115] Feature code unit: used for converting the best face image into a feature code in the central server, and transmitting the feature code to a preset comparison server;

[0116] Similarity calculation unit: used to calculate and compare the feature code with the pre-recorded face model in the comparison server, determine the similarity score, and transmit the similarity score to the security identification center.

[0117] The working principle and beneficial effects of the above technical solution are:

[0118] In this technical solution, the best face image unit is used to evaluate the face image quality of the same face image set and output the best face image; the best face image is an image with a frontal posture and covers the coordinates of all face feature points, so as to improve the accurate recognition of face feature points; the feature code unit is used to convert the best face image into a feature code in the central server and transmit the feature code to the preset comparison server; the similarity calculation unit is used to calculate and compare the feature code with the pre-recorded face model in the comparison server, determine the similarity score, and transmit the similarity score to the security identification center; the feature point acquisition unit is used to input all the face images to be processed of the same face and extract the coordinates of the standard face feature points; the calculation unit is used to calculate the constant matrix based on the coordinates of the standard face feature points and calculate the quality numerical results of all the face images to be processed of the same face; the quality evaluation unit is used to sort the quality numerical results, determine the sorting results, and output the best face image based on the sorting results. The data quality of the face image is improved, the resolution of the image is clearer and more variable, and the fault tolerance rate is improved.

[0119] Embodiment 6:

[0120] In one embodiment, the facial feature point coordinates include at least left eye coordinates, right eye coordinates, nose coordinates, left mouth corner coordinates, and right mouth corner coordinates.

[0121] Embodiment 7:

[0122] In one embodiment, the best face image is an image with a frontal posture and covering the coordinates of all facial feature points.

[0123] Embodiment 8:

[0124] In one embodiment, the security identification module includes:

[0125] Collection unit: used to receive the similarity score, display the data and generate a report;

[0126] The judgment and recognition unit is used to compare and judge the report with a preset threshold, determine whether the optimal face image has a security anomaly, and obtain a recognition result; wherein,

[0127] When the similarity score is less than a preset threshold, the recognition result is recognition failure;

[0128] When the similarity score is greater than a preset threshold, the recognition result is a successful recognition;

[0129] Result management unit: used to respond to access control actions based on the identification results and implement relevant management solutions.

[0130] The working principle and beneficial effects of the above technical solution are:

[0131] In this technical solution, the security identification module includes: a collection unit for receiving similarity scores, displaying data, and generating reports; a judgment and identification unit for comparing and judging the reports with preset thresholds, determining whether the best face image has security anomalies, and obtaining recognition results; when the similarity score is less than the preset threshold, the recognition result is recognition failure; when the similarity score is greater than the preset threshold, the recognition result is recognition success; a result management unit for responding to access control actions based on the recognition results and implementing relevant management solutions. Through the security identification module, relevant management solutions are implemented on the basis of face management and recognition.

[0132] Embodiment 9:

[0133] In one embodiment, the result management unit includes:

[0134] Successful identification management subunit: when the identification result is successful, the permission is granted, the corresponding door is controlled to open, and the face image is reversely propagated to restore it to the same face image set, and the capacity of the face image sample library is supplemented;

[0135] Identification failure management subunit: When the recognition result is recognition failure, no permission is granted, the corresponding access control is controlled to implement the prohibition action, and the best face image is reversely propagated to restore it to the same face image set, and the real-time position is tracked and locked and transmitted to the security terminal.

[0136] The working principle and beneficial effects of the above technical solution are:

[0137] In this technical solution, the recognition success management subunit: when the recognition result is recognition success, the permission is granted, and the corresponding access control is controlled to perform the opening action, and the best face image is back-propagated to restore it to the same face image set, and the capacity of the face image sample library is supplemented; the recognition failure management subunit: when the recognition result is recognition failure, the permission is not granted, and the corresponding access control is controlled to perform the prohibition action, and the best face image is back-propagated to restore it to the same face image set, and the real-time position is tracked and locked, and transmitted to the security terminal. Through the correct recognition of the face, the user's security identification and security management are improved.

[0138] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A face recognition system based on digital image processing technology, characterized in that: include: Control area module: divides the control area by a preset attribute threshold value to determine the corresponding target range; wherein the attribute threshold value represents the attribute threshold value of the display attribute of the control area; Face data acquisition module: collects face images through image sensors preset within the target range, and establishes a face image sample library based on the face images; Intelligent environment control module: detects the face image sample library, extracts the face point coordinates of the face image to be identified through feature training, and determines the same face image set; Comparison server module: optimally output the same face image set, process it using the comparison server, generate a similarity score, and transmit the similarity score to a preset security identification center; Security identification module: used to collect the similarity scores, determine the identification results through a preset threshold, and form a corresponding management plan; Wherein, the control area module includes: Area setting unit: used to set up an integrated control system in the preset terminal equipment, form an integrated range in the control area, and build a data sharing platform; Parameter acquisition unit: used to acquire environmental parameters within the integration range, upload them to the data sharing platform, identify their display attributes, and generate a parameter catalog; wherein, The display attributes include at least texture attributes, object attributes and detail attributes within the integration range; Control area unit: used to adapt the parameter catalog according to the preset attribute threshold value and determine the corresponding target range through segmentation; wherein, The control area at least includes a cross control area and a separation control area.

2. A face recognition system based on digital image processing technology as claimed in claim 1, characterized in that: The face data acquisition module comprises: A face acquisition unit is used to acquire a real-time image containing face features in the control area through an image sensor preset in the control area; wherein: The image sensor includes at least one or more of a face recognition camera, a face recognition camera, a face recognition gate and a PDA data collector; Image pre-screening unit: used to pre-screen the real-time image based on pre-processing technology, remove images with insufficient facial features in the real-time image, and determine the facial image to be processed; wherein, The preprocessing technology at least includes geometric correction, image filtering, image enhancement and image edge detection methods; Image sample library unit: used for transmitting the face image to be processed to the data sharing platform and integrating it into a face image sample library.

3. A face recognition system based on digital image processing technology as claimed in claim 1, characterized in that: The intelligent environmental control module comprises: Face detection unit: used for inputting the face image sample library into a preset detection network to obtain the position of the face candidate frame in the face image to be processed; Feature training unit: used for performing convolutional neural network learning on the position of the face candidate frame to determine the coordinates of the face feature points; The same face unit is used to process the face image to be processed based on the coordinates of the face feature points by applying a classification method to form a set of the same face images; wherein, The classification method at least includes a maximum likelihood classification method and a minimum distance classification method; The feature training unit comprises: The first convolutional network subunit is used for capturing the face candidate frame image unit based on the position of the face candidate frame to generate a candidate frame screenshot; The second convolutional network subunit is used to integrate the candidate frame screenshots, build a face recognition network based on the face image sample library, and adjust the parameters of the face recognition network to build a face recognition model; Feature point extraction subunit: input the face image to be processed into the face recognition model to obtain the coordinates of the face feature points.

4. A face recognition system based on digital image processing technology as claimed in claim 1, characterized in that: The comparison server module comprises: Feature point acquisition unit: used to acquire all the facial images to be processed of the same face and extract the coordinates of standard facial feature points; A calculation unit is used to obtain pixel points of the face image to be processed, project each pixel point on the face image to be processed into a preset parameter domain, and obtain the filtered pixel points; ； in Represents the first face image to be processed after screening Line Column pixels; Representative Line The pixel value of the pixel point in the column, Represents the preset parameter domain for face images, represents the inverse of the preset parameter domain for face images, Representative Line The pixel value after the pixel point of the column is transposed, Represents the preset parameter domain for face images The ideal range of values ​​for the rows, Represents the preset parameter domain for face images The ideal range of values ​​for the column; Accumulated peak unit: used to accumulate the pixel points of the face image to be processed through a preset sampling period and calculate the accumulated peak value; ； in, represents the cumulative peak value, represent No. The sampling period of the pixels in a row, represent No. The sampling period of the pixel points of the column; Standard facial feature point coordinate unit: used to analyze the points corresponding to the accumulated peak values ​​and map them to the facial image parameter space to obtain the standard facial feature point coordinates; Constant matrix unit: used to constantize the coordinates of the standard facial feature points and calculate the constant matrix of the coordinates of the standard facial feature points; A quality numerical result unit is used to estimate the parameters of the constant matrix of all the facial images to be processed of the same face by using a preset quality evaluation index, and determine the quality numerical results of all the facial images to be processed of the same face; A quality evaluation unit: used for sorting the quality value results, determining the sorting result, and outputting the best face image based on the sorting result; Feature code unit: used for converting the best face image into a feature code in the central server, and transmitting the feature code to a preset comparison server; Similarity calculation unit: used to calculate and compare the feature code with the pre-recorded face model in the comparison server, determine the similarity score, and transmit the similarity score to the security identification center.

5. A face recognition system based on digital image processing technology as claimed in claim 3, characterized in that: The facial feature point coordinates include at least left eye coordinates, right eye coordinates, nose coordinates, left mouth corner coordinates and right mouth corner coordinates.

6. A face recognition system based on digital image processing technology as claimed in claim 4, characterized in that: The best face image is an image with a frontal posture and covering the coordinates of all facial feature points.

7. A face recognition system based on digital image processing technology as claimed in claim 1, characterized in that: The security identification module comprises: Collection unit: used to receive the similarity score, display the data and generate a report; The judgment and recognition unit is used to compare and judge the report with a preset threshold, determine whether the best face image has security anomalies, and obtain the recognition result; wherein, When the similarity score is less than a preset threshold, the recognition result is recognition failure; When the similarity score is greater than a preset threshold, the recognition result is a successful recognition; Result management unit: used to respond to access control actions based on the identification results and implement relevant management solutions.

8. A face recognition system based on digital image processing technology as claimed in claim 7, characterized in that: The result management unit comprises: Successful identification management subunit: when the identification result is successful, the permission is granted, the corresponding door is controlled to open, and the face image is reversely propagated to restore it to the same face image set, and the capacity of the face image sample library is supplemented; Identification failure management subunit: When the recognition result is recognition failure, no permission is granted, the corresponding access control is controlled to implement the prohibition action, and the best face image is reversely propagated to restore it to the same face image set, and the real-time position is tracked and locked and transmitted to the security terminal.

Images