Face acquisition and verification system and method

Abstract

The invention belongs to the technical field of face collection, particularly discloses a face collection and verification system and method, and aims to solve the problems that an existing face collection and verification method is tedious in process, low in efficiency and prone to causing conflict emotion of a collected person. Complete certificate information in an identity card is directly read through the mobile phone NFC module, meanwhile, the video recorder continuously records videos and transmits the videos to the video cloud platform in real time, original step-by-step operation is integrated into one-step linkage, and the verification amount in unit time is greatly improved. And then, the system automatically screens data of'registered identities but not collected faces' through a timed task, positions a corresponding video time period through a space-time task, calls a face retrieval service to complete'searching images by images' and intercepts an optimal face, and automatically completes identity verification and information input from a video based on space-time information, so that the purposes of rapidness and high efficiency are achieved.

Description

Technical Field

[0001] This invention relates to the field of face capture technology, specifically to a face capture and verification system and method. Background Technology

[0002] With the continuous improvement of public service scenario management requirements, facial recognition technology has become a core support means for personnel identity verification, and is widely used in various scenarios such as security for large-scale events and registration for exhibition entry. This technology can quickly complete identity verification by associating facial biometric features with document information, thus optimizing the passage efficiency in public service scenarios.

[0003] Currently, when operators verify identities, they use a mobile identity verification app to complete the operation in two steps. The first step is to manually enter the ID number, name, and other document information of the person being collected. The second step is to manually switch to the app's shooting mode and take a photo of the person's face to complete the biometric data collection.

[0004] However, the above-mentioned methods have limitations in actual use. On the one hand, a single data collection process involves multiple steps and takes a long time. During peak hours, it can easily cause on-site queuing and congestion, resulting in low overall verification efficiency and making it difficult to meet the needs of rapid passage in high-traffic scenarios. On the other hand, multiple manual operations will prolong the interaction time between the operator and the data subject, which not only increases the workload of the operator but also makes the data subject feel resistant due to the long waiting time. Summary of the Invention

[0005] The purpose of this invention is to provide a face collection and verification system and method to solve the problems of existing face collection and verification methods, which are cumbersome, inefficient, and prone to causing resistance from those being collected.

[0006] To achieve the above objectives, the basic solution provided by this invention is as follows: a face acquisition and verification system, comprising a mobile phone NFC module, a personnel information database, a Redis database ordered set Zset module, a video recorder, a video cloud platform, and a face retrieval service module. The NFC module communicates with the personnel information database via the HTTP protocol, the personnel information database communicates with the Redis database ordered set Zset module via the HTTP protocol, the video recorder communicates with the video cloud platform via the HTTP protocol, the video cloud platform communicates with the face retrieval service module via the HTTP protocol, and the face retrieval service module communicates with the personnel information database via the HTTP protocol.

[0007] The beneficial effects of this invention are as follows: It directly reads complete document information from an ID card via a mobile phone's NFC module, while a video recorder continuously records video and transmits it in real-time to a video cloud platform, integrating the original step-by-step operations into a single, coordinated process, significantly increasing the verification volume per unit time. Next, the personnel information database filters data for "registered identities without facial recognition," and then uses a facial retrieval service to capture the optimal face. Based on spatiotemporal information, identity verification and information entry are automatically completed from the video, achieving a fast and efficient process.

[0008] Option 2, a method for facial recognition and verification, includes the following steps: S1. First, the operator uses a mobile phone with an app to approach the person's ID card and reads the ID photo and ID number through the phone's NFC module. The operator then sends the read ID photo and ID number to the backend. After receiving the data, the backend stores it in the personnel information database. At the same time, the video recorder worn by the operator continuously records video. The video is transmitted to the video cloud platform in real time and the facial photo information is analyzed. S2. The personnel information database has stored the information ID of the person being collected, the entry time, the verification mode, the ID number, the ID photo, the collector information, and the video recorder ID bound to the collector. At present, the two fields of "collected photo" and "similarity" in the personnel information database have no stored content and are in a state of waiting to be stored. S3. Next, the personnel information database queries the data of "verification mode and ID photo have stored content, but the collected photo has no stored content" every minute, and stores the queried data into the sorted set Zset module of the Redis database; S4. Then, every 30 seconds, the personnel information database executes the Redis database sorted set Zset module to query the data in the time period from 270 to 300 seconds before the current time, and puts the collected person information ID in the retrieved data into a Java queue. The Java queue queries the personnel information database in turn using the collected person information ID. After finding the collected person information ID, it locates the time period of the video that captured the face in the video cloud platform based on the database entry time related to the collected person information ID and the video recorder ID bound to the collected person. S5. Subsequently, the personnel information database sends a request to the face retrieval service, submitting an ID photo, time range, and video recorder ID, requesting the "image search" service in the video cloud platform. The face retrieval service extracts the most similar and best face image from the video and returns it to the personnel information database. The personnel information database updates the returned collected photos and similarity, thus completing a personnel information record.

[0009] Option 3, which is the preferred option of Option 2, involves transmitting the video recorded by the video recorder to the video cloud platform in S1, and then querying it after a 5-minute delay. The code does not process the task immediately upon receiving it, but instead makes each task wait 5 minutes before execution, ensuring that when the system captures a face, the corresponding video segment has already been stored in the cloud, thus avoiding retrieval failures before the video has arrived.

[0010] Option 4, which is the preferred option of Option 2, involves storing the collected person's information ID as the value and the entry time as the score in the Redis database's sorted set Zset module in S3. Attached Figure Description

[0011] Figure 1 This is a flowchart of a face capture and verification system and method according to the present invention. Detailed Implementation

[0012] The present invention will be further described in detail below through specific embodiments: Example 1 like Figure 1 The system described above is a face capture and verification system, comprising a mobile phone NFC module, a personnel information database, a Redis database sorted set (Zset) module, a video recorder, a video cloud platform, and a face retrieval service module. The NFC module communicates with the personnel information database via the HTTP protocol, as do the personnel information database and the Redis database sorted set (Zset) module, the video recorder and the video cloud platform, the video cloud platform and the face retrieval service module, and the face retrieval service module and the personnel information database.

[0013] Example 2 like Figure 1 As shown: A face capture and verification system includes the following steps: S1. First, the operator uses a mobile phone with an app (for identity verification) to approach the person's ID card. The operator reads the ID photo and ID number through the phone's NFC module. The operator then sends the read ID photo and ID number to the backend. After receiving the data, the backend stores it in the personnel information database. At the same time, the video recorder worn by the operator continuously records video. The video is transmitted to the video cloud platform in real time and the facial photo information is analyzed. After the video recorded by the video recorder is transmitted to the video cloud platform, it is queried after a 5-minute delay. The code does not process the task immediately after receiving it, but makes each task wait for 5 minutes before execution. This ensures that when the system captures the face, the corresponding video segment has already been stored in the cloud, avoiding retrieval failures when the video has not yet arrived. S2. The personnel information database has stored the information ID of the person being collected, the entry time, the verification mode, the ID number, the ID photo, the collector information, and the video recorder ID bound to the collector. At present, the two fields of "collected photo" and "similarity" in the personnel information database have no stored content and are in a state of waiting to be stored. S3. Next, the personnel information database queries the data of "verification mode and ID photo have stored content, but the collected photo has no stored content" every minute, and stores the queried data into the sorted set Zset module of the Redis database. The ID of the person whose information is collected is used as the value, and the entry time is used as the score. S4. Then, every 30 seconds, the personnel information database executes the Redis database sorted set Zset module to query the data in the time period from 270 to 300 seconds before the current time, and puts the collected person information ID in the retrieved data into a Java queue. The Java queue queries the personnel information database in turn using the collected person information ID. After finding the collected person information ID, it locates the time period of the video that captured the face in the video cloud platform based on the database entry time related to the collected person information ID and the video recorder ID bound to the collected person. S5. Subsequently, the personnel information database sends a request to the face retrieval service, submitting an ID photo, time range, and video recorder ID, requesting the "image search" service in the video cloud platform. The face retrieval service extracts the most similar and best face image from the video and returns it to the personnel information database. The personnel information database updates the returned collected photos and similarity, thus completing a personnel information record.

[0014] The above descriptions are merely embodiments of the present invention, and common knowledge regarding specific structures and characteristics is not elaborated upon here. It should be noted that those skilled in the art can make various modifications and improvements without departing from the structure of the present invention, and these should also be considered within the scope of protection of the present invention. These modifications and improvements will not affect the effectiveness of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A face capture and verification system, characterized in that, The system includes a mobile phone NFC module, a personnel information database, a Redis database sorted set (Zset) module, a video recorder, a video cloud platform, and a face retrieval service module. The NFC module communicates with the personnel information database via the HTTP protocol, as do the personnel information database and the Redis database sorted set (Zset) module, the video recorder and the video cloud platform, the video cloud platform and the face retrieval service module, and the face retrieval service module and the personnel information database.

2. A method for face capture and verification, characterized in that, Includes the following steps: S1. First, the operator uses a mobile phone with an app to approach the person's ID card and reads the ID photo and ID number through the phone's NFC module. The operator then sends the read ID photo and ID number to the backend. After receiving the data, the backend stores it in the personnel information database. At the same time, the video recorder worn by the operator continuously records video. The video is transmitted to the video cloud platform in real time and the facial photo information is analyzed. S2. The personnel information database has stored the information ID of the person being collected, the entry time, the verification mode, the ID number, the ID photo, the collector information, and the video recorder ID bound to the collector. At present, the two fields of "collected photo" and "similarity" in the personnel information database have no stored content and are in a state of waiting to be stored. S3. Next, the personnel information database queries the data of "verification mode and ID photo have stored content, but the collected photo has no stored content" every minute, and stores the queried data into the Redis database sorted set Zset module; S4. Then, every 30 seconds, the personnel information database executes the Redis database sorted set Zset module to query the data in the time period from 270 to 300 seconds before the current time, and puts the collected person information ID in the retrieved data into a Java queue. The Java queue queries the personnel information database in turn using the collected person information ID. After finding the collected person information ID, it locates the time period of the video that captured the face in the video cloud platform based on the database entry time related to the collected person information ID and the video recorder ID bound to the collected person. S5. Subsequently, the personnel information database sends a request to the face retrieval service, submitting an ID photo, time range, and video recorder ID, requesting the "image search" service in the video cloud platform. The face retrieval service extracts the most similar and best face image from the video and returns it to the personnel information database. The personnel information database updates the returned collected photos and similarity, thus completing a personnel information record.

3. The face capture and verification method according to claim 2, characterized in that, In S1, the video recorded by the video recorder is transmitted to the video cloud platform and can be queried after a 5-minute delay.

4. The face capture and verification method according to claim 2, characterized in that, In S3, the ID of the person whose information is being collected is used as the value, and the entry time is used as the score. These are stored in the sorted set Zset module of the Redis database.

Images