Face authentication method based on target group

Abstract

The invention discloses a face authentication method based on a target group. The method comprises the following steps that S1, a cloud integrates users with input user feature codes to form an emergency contact group; s2, the cloud sends the user position data and the call reminding information to all other users in the emergency contact group, or the cloud sends the image data subjected to face recognition, the call reminding information and the user position data to all other users in the emergency contact group; by the adoption of the method, face recognition is carried out when a call is made, if face recognition succeeds, other users in the emergency contact group can know who makes a call with the user and the frequency of call communication with the user, and therefore the other users can prevent telecommunication fraud.

Description

Technical Field

[0001] The present invention relates to the technical field of control and regulation systems, and in particular to a face recognition method based on a target group. Background Art

[0002] With the rapid development of mobile communications and Internet technologies, people's daily lives are becoming more and more dependent on smartphones. However, today's telecommunications fraud methods are endless, and children and the elderly at home are easily deceived.

[0003] In the prior art, although there are some applications that can help family members stay in touch in an emergency by sending location information or emergency distress signals, these applications usually require users to manually trigger them, or the operations are cumbersome, and these operations are too complicated for children and the elderly.

[0004] Therefore, there is a need for a system that can automatically identify the user's emergency status and take corresponding measures according to different situations, so as to effectively improve the safety of family members in emergency situations. Summary of the invention

[0005] In view of this, the purpose of the present invention is to provide a target group-based facial authentication method, which performs facial recognition when making a phone call so that other family members can know who the user is talking to, thereby effectively preventing telecommunications fraud, or when someone steals a mobile phone, other family members can call the user and see the thief's face.

[0006] The technical solution adopted by the present invention to solve the technical problem is to provide a target group-based face recognition method, including a cloud end and a user end, wherein the user end includes a plurality of positioning modules corresponding to the user one by one, a plurality of camera modules corresponding to the user one by one, and a plurality of control modules corresponding to the user one by one, and the cloud end includes a server;

[0007] The method comprises the following steps:

[0008] S1. The control module performs face recognition on the user through the camera module to obtain the user's facial image data, and the control module generates a unique user feature code corresponding to the user's facial image data. When the user inputs at least two user feature codes, the control module with the input user feature codes transmits the user's facial image data to the cloud, and the cloud integrates the users with the input user feature codes to form an emergency contact group;

[0009] S2. When a call is made to the user or the user makes an outgoing call, the control module obtains the user's location data through the positioning module, and performs face recognition on the user through the camera module. If the face recognition result matches the user's face image data, the control module will send a call reminder message and the user's location data to the cloud, and the cloud will send the user's location data and call reminder message to all other users in the emergency contact group;

[0010] If the face recognition result does not match the user's facial image data, the control module will send the face recognition image data, phone reminder information and user location data to the cloud, and the cloud will send the face recognition image data, phone reminder information and user location data to all other users in the emergency contact group.

[0011] As a further improvement of the present invention, the telephone reminder information includes the user's own telephone number, the other party's telephone number, the time of making the call and the type of the call, and the type of the call is the user being called or the user making the call.

[0012] As a further improvement of the present invention, in step S2, the cloud sends the user location data and phone reminder information to all other users in the emergency contact group, and also includes the following steps: if there are two different phone reminder messages in the cloud, and the other party number in one phone reminder message is equal to the user's own phone number in the other phone reminder message, and there is no image data in the cloud, then it is determined that the two users in the emergency contact group are calling each other at this time, and the cloud stops sending information.

[0013] As a further improvement of the present invention, in step S1, the control module input with the user feature code transmits the user's facial image data to the cloud while the control module transmits the user's own phone number to the cloud, and the cloud forms the user's standard phone number according to the user's own phone number;

[0014] In step S2, the user makes a call specifically by dialing or directly making a call without dialing. When the user makes a call directly without dialing, the user directly dials the standard telephone numbers of other users in the emergency contact group.

[0015] As a further improvement of the present invention, the specific steps of the control module in step S2 sending the image data for face recognition to the cloud include:

[0016] S21, the control module reads the video file for face recognition, and pre-processes the video file to obtain a pre-processed video file;

[0017] S22, the control module locates the face region of each frame in the preprocessed video file by using a face detection algorithm, locates multiple key points in the face region by using a key point detection algorithm, selects the frame with the largest number of key points and integrates them to form a key frame set;

[0018] S23, the control module is preset with a standard definition, performs image definition evaluation on a key picture set, selects pictures with a definition higher than the standard definition and performs approximate picture deletion processing, and transmits the remaining pictures to the cloud.

[0019] As a further improvement of the present invention, the specific steps of preprocessing the video file to obtain the preprocessed video file in step S21 are: performing brightness preprocessing on the video file to obtain the preprocessed video file.

[0020] As a further improvement of the present invention, there are five key points, which are respectively the left eye, the right eye, the nose, the mouth, the left ear and the right ear.

[0021] As a further improvement of the present invention, the specific steps of locating the face region of each frame in the preprocessed video file by the control module through the face detection algorithm in step S22 are: the control module identifies and locates the face region in each frame through the Haar classifier;

[0022] The specific steps of locating the multiple key points in the face area by using the key point detection algorithm in step S22 are: locating the multiple key points in the located face area by using the shape predictor in Dlib.

[0023] As a further improvement of the present invention, the specific step of performing image definition evaluation on the key picture set in step S23 is: performing image definition evaluation on each picture in the key picture set by using the Laplacian variance Laplace zoom method.

[0024] As a further improvement of the present invention, the specific steps of the approximate picture deletion processing in step S23 are: use the FLANN feature matching algorithm to match the feature points between the two adjacent pictures, and when the matching score is higher than the standard matching score preset in the control module, delete the previous picture or delete the next picture.

[0025] The beneficial effects of the present invention are at least as follows: the method performs face recognition when making a phone call. If the face recognition is successful, other users in the emergency contact group can know who the user is calling and the frequency of the phone communication, so that other users can prevent telecommunications fraud; when the face recognition fails, the image data for face recognition is sent so that other users in the emergency contact group can see the picture, so as to determine whether the user is in danger. After someone steals the mobile phone, other family members can see the face of the thief as long as they call the user. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0027] The technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings.

[0028] Reference Figure 1 , the present invention proposes a target group-based face recognition method, including a cloud end and a user end, the user end includes a plurality of positioning modules corresponding to the user one by one, a plurality of camera modules corresponding to the user one by one, and a plurality of control modules corresponding to the user one by one, the cloud end includes a server;

[0029] The method comprises the following steps:

[0030] S1. The control module performs face recognition on the user through the camera module to obtain the user's facial image data, and the control module generates a unique user feature code corresponding to the user's facial image data. When the user inputs at least two user feature codes, the control module with the input user feature codes transmits the user's facial image data to the cloud, and the cloud integrates the users with the input user feature codes to form an emergency contact group;

[0031] S2. When a call is made to the user or the user makes an outgoing call, the control module obtains the user's location data through the positioning module, and performs face recognition on the user through the camera module. If the face recognition result matches the user's face image data, the control module will send a call reminder message and the user's location data to the cloud, and the cloud will send the user's location data and call reminder message to all other users in the emergency contact group;

[0032] If the face recognition result does not match the user's facial image data, the control module will send the face recognition image data, phone reminder information and user location data to the cloud, and the cloud will send the face recognition image data, phone reminder information and user location data to all other users in the emergency contact group.

[0033] This method performs face recognition when making a phone call. If the face recognition is successful, other users in the emergency contact group can know who the user is talking to and the frequency of the phone communication, so that other users can prevent telecommunications fraud; when the face recognition fails, sending the image data for face recognition can allow other users in the emergency contact group to see the picture, so as to judge whether the user is in danger; after someone steals the mobile phone, other family members can see the thief's face as long as they call the user. Attention should be paid at this time. It is not necessary to answer the call to perform face recognition. Otherwise, if the face recognition is performed after the call is answered, the mobile phone generally cannot capture the face because it is placed next to the ear. Therefore, if the user moves very quickly and the face cannot be captured, other users in the emergency contact group can also judge whether the user has placed the mobile phone next to the ear, resulting in the failure of face recognition, by observing the image data for face recognition.

[0034] As a further improvement of the present invention, the telephone reminder information includes the user's own telephone number, the other party's telephone number, the time of making the call and the type of the call, and the type of the call is the user being called or the user making the call.

[0035] As a further improvement of the present invention, in step S2, the cloud sends the user location data and phone reminder information to all other users in the emergency contact group, and also includes the following steps: if there are two different phone reminder messages in the cloud, and the other party number in one phone reminder message is equal to the user's own phone number in the other phone reminder message, and there is no image data in the cloud, then it is determined that the two users in the emergency contact group are calling each other at this time, and the cloud stops sending information.

[0036] In this case, if two users in the same emergency contact group call outsiders at the same time, it will not be judged as the two users calling each other; it should be noted that the premise of this step is that both parties have passed the face recognition. If one party fails the face recognition, it is possible that the life of one user is in danger, and the image data needs to be uploaded to the cloud and sent to all other users; it should be noted that because current mobile phones have dual SIM cards and dual standby, it is possible that the user dials through a second mobile phone number, but as long as the mobile phone is not changed and the user's face is not changed, this method can be used to stop the cloud from sending information to avoid redundant information.

[0037] Another great advantage of this step method is that it can alarm silently. The existing emergency alarm is through long pressing the function key and swiping the screen. When the user is about to be in danger of life, it is not convenient to call home directly to say that he is in danger, otherwise it may directly arouse the anger of the criminals. The user can deliberately prevent the phone from capturing the face, shake the phone so that the phone can capture the criminals, let the family know and call the police; similarly, if the criminals kidnap the user and use the user's mobile phone to call anyone to communicate, the criminals' facial information can be transmitted silently to other users, or when the criminals ask the user to call their family for ransom, the criminals' facial information can be transmitted silently to other users without the criminals knowing.

[0038] As a further improvement of the present invention, in step S1, the control module input with the user feature code transmits the user's facial image data to the cloud while the control module transmits the user's own phone number to the cloud, and the cloud forms the user's standard phone number according to the user's own phone number;

[0039] In step S2, the user makes a call specifically by dialing or directly making a call without dialing. When the user makes a call directly without dialing, the user directly dials the standard telephone numbers of other users in the emergency contact group.

[0040] This method can be used to quickly call your family. For example, if you are in danger and cannot dial, you can dial an empty number to quickly call your family.

[0041] As a further improvement of the present invention, the specific steps of the control module in step S2 sending the image data for face recognition to the cloud include:

[0042] S21, the control module reads the video file for face recognition, and pre-processes the video file to obtain a pre-processed video file;

[0043] S22, the control module locates the face region of each frame in the preprocessed video file by using a face detection algorithm, locates multiple key points in the face region by using a key point detection algorithm, selects the frame with the largest number of key points and integrates them to form a key frame set;

[0044] S23, the control module is preset with a standard definition, performs image definition evaluation on a key picture set, selects pictures with a definition higher than the standard definition and performs approximate picture deletion processing, and transmits the remaining pictures to the cloud.

[0045] As a further improvement of the present invention, the specific steps of preprocessing the video file to obtain the preprocessed video file in step S21 are: performing brightness preprocessing on the video file to obtain the preprocessed video file.

[0046] As a further improvement of the present invention, there are five key points, which are respectively the left eye, the right eye, the nose, the mouth, the left ear and the right ear.

[0047] As a further improvement of the present invention, the specific steps of locating the face region of each frame in the preprocessed video file by the control module through the face detection algorithm in step S22 are: the control module identifies and locates the face region in each frame through the Haar classifier;

[0048] The specific steps of locating the multiple key points in the face area by using the key point detection algorithm in step S22 are: locating the multiple key points in the located face area by using the shape predictor in Dlib.

[0049] As a further improvement of the present invention, the specific step of performing image definition evaluation on the key picture set in step S23 is: performing image definition evaluation on each picture in the key picture set by using the Laplacian variance Laplace zoom method.

[0050] As a further improvement of the present invention, the specific steps of the approximate picture deletion processing in step S23 are: use the FLANN feature matching algorithm to match the feature points between the two adjacent pictures, and when the matching score is higher than the standard matching score preset in the control module, delete the previous picture or delete the next picture.

Claims

1. A target group-based face recognition method, characterized in that: It includes a cloud end and a user end, wherein the user end includes a plurality of positioning modules corresponding to the users one by one, a plurality of camera modules corresponding to the users one by one, and a plurality of control modules corresponding to the users one by one, and the cloud end includes a server; The method comprises the following steps: S1. The control module performs face recognition on the user through the camera module to obtain the user's facial image data, and the control module generates a unique user feature code corresponding to the user's facial image data. When the user inputs at least two user feature codes, the control module with the input user feature codes transmits the user's facial image data to the cloud, and the cloud integrates the users with the input user feature codes to form an emergency contact group; S2. When a call is made to the user or the user makes an outgoing call, the control module obtains the user's location data through the positioning module, and performs face recognition on the user through the camera module. If the face recognition result matches the user's face image data, the control module will send a call reminder message and the user's location data to the cloud, and the cloud will send the user's location data and call reminder message to all other users in the emergency contact group; If the face recognition result does not match the user's facial image data, the control module will send the face recognition image data, phone reminder information and user location data to the cloud, and the cloud will send the face recognition image data, phone reminder information and user location data to all other users in the emergency contact group.

2. The target group-based face authentication method according to claim 1, characterized in that: The telephone reminder information includes the user's own telephone number, the other party's telephone number, the time of the call and the type of call, and the type of call is the user being called or the user making the call.

3. The target group-based face authentication method according to claim 1, characterized in that: In step S2, the cloud sends the user location data and phone reminder information to all other users in the emergency contact group, and also includes the following steps: if there are two different phone reminder messages in the cloud, and the other party number in one phone reminder message is equal to the user's own phone number in the other phone reminder message, and there is no image data in the cloud, it is determined that the two users in the emergency contact group are calling each other at this time, and the cloud stops sending information.

4. The target group-based face authentication method according to claim 1, characterized in that: In step S1, the control module inputs the user feature code and transmits the user's facial image data to the cloud. At the same time, the control module transmits the user's own phone number to the cloud, and the cloud forms the user's standard phone number based on the user's own phone number; In step S2, the user makes a call specifically by dialing or directly making a call without dialing. When the user makes a call directly without dialing, the user directly dials the standard telephone numbers of other users in the emergency contact group.

5. The target group-based face authentication method according to claim 1, characterized in that: In step S2, the control module sends the image data for face recognition to the cloud in the following specific steps: S21, the control module reads the video file for face recognition, and pre-processes the video file to obtain a pre-processed video file; S22, the control module locates the face region of each frame in the preprocessed video file by using a face detection algorithm, locates multiple key points in the face region by using a key point detection algorithm, selects the frame with the largest number of key points and integrates them to form a key frame set; S23, the control module is preset with a standard definition, performs image definition evaluation on a key picture set, selects pictures with a definition higher than the standard definition and performs approximate picture deletion processing, and transmits the remaining pictures to the cloud.

6. The target group-based face authentication method according to claim 5, characterized in that: The specific steps of preprocessing the video file to obtain the preprocessed video file in step S21 are: performing brightness preprocessing on the video file to obtain the preprocessed video file.

7. The target group-based face authentication method according to claim 5, characterized in that: There are five key points, which are the left eye, the right eye, the nose, the mouth, the left ear and the right ear.

8. The target group-based face authentication method according to claim 5, characterized in that: In step S22, the control module locates the face region of each frame in the preprocessed video file by using the face detection algorithm. The control module identifies and locates the face region in each frame by using the Haar classifier. The specific steps of locating the multiple key points in the face area by using the key point detection algorithm in step S22 are: locating the multiple key points in the located face area by using the shape predictor in Dlib.

9. The target group-based face authentication method according to claim 5, characterized in that: The specific steps of performing image definition evaluation on the key picture set in step S23 are: performing image definition evaluation on each picture in the key picture set by using the Laplacian variance Laplace zoom method.

10. The target group-based face authentication method according to claim 5, characterized in that: The specific steps of the approximate picture deletion processing in step S23 are: use the FLANN feature matching algorithm to match the feature points between the two adjacent pictures, and when the matching score is higher than the standard matching score preset in the control module, delete the previous picture or delete the next picture.

Images