System and method for judging face visibility of ATM user

Abstract

The invention discloses an automatic teller machine user face visibility judging system and method. The system comprises a camera installed above a screen of an automatic teller machine and used for taking pictures of human faces outwards, a camera installed on the top of the automatic teller machine and used for taking pictures of an entrance of the automatic teller machine downwards, an analyzing module installed in a shell body of the automatic teller machine, and a monitoring center connected with the analyzing module through the network. According to the method, the fact whether the automatic teller machine is used by a person is judged by means of image outline structure comparison and pixel region comparison; when the fact that the automatic teller machine is used by a person is confirmed, the face and facial organs of the person are detected by a cascading AdaBoost detector in the area of another path of video center; the fact whether the face of the person is blocked is judged by means of an eye classifier and a mouth classifier which are based on a Gaussian kernel support vector machine (SVM); the result in a time window is analyzed to obtain a result; the analyzing module transmits the result to the monitoring center through a private network or a wide area network for timely warning and recording and archiving.

Description

technical field [0001] The invention belongs to the fields of computer vision and artificial intelligence, and in particular relates to a system and method for judging the face visibility of an automatic teller machine user. Background technique [0002] In order to prevent various criminal activities using ATM machines, such as stealing other people's bank cards, inducing others to transfer deposits to an account, and then withdrawing them on the ATM machine, etc., most of the existing ATM machines have been installed with camera monitoring The system captures the face of the ATM machine user, and searches for the suspected criminal withdrawal operation and the withdrawal person based on the monitoring data after the case occurs. However, criminal suspects often use various methods to cover their faces to hide their identities, which makes it difficult to solve the case. Take snapshots of users who do not cover their faces, give warnings to users with covered faces in a ti...

AI Technical Summary

Problems solved by technology

However, faces in the real world have different shapes, and there may always be some weird faces that cause the judgment to fail. Therefore, there are also problems with the application layer logic that prevents users from using it.

Images