Living body recognition method, storage medium, and computer device
摘要
說(shuō)明書
Specifically, both the first confidence level and the second confidence level are confidence levels of the target image being a live facial image, and are confidence levels obtained through analysis based on different image features. Therefore, the server may integrate the two confidence levels to obtain a final confidence level, and obtain, according to the final confidence level, a recognition result indicating whether the target image is a live facial image. That is, the server may perform a score fusion to fuse the two scores or confidence levels so as to obtain a final score or confidence level.
Further, in an authentication scene, after obtaining a recognition result indicating whether the target image is a live facial image, the server can obtain, according to this recognition result and the facial recognition result, an authentication result indicating whether the authentication succeeds, and perform operations corresponding to the authentication result. This can ensure that the operations are performed by the user himself/herself. For example, in a process of opening a bank account in a bank application, if it is determined that the target image is a live facial image and the facial recognition indicates matching, the authentication succeeds and subsequent account opening operations go on. For another example, in an access control scene, if it is determined that the target image is a live facial image and the facial recognition indicates matching, the authentication succeeds and a door opening instruction is output.
