H5 video active detection face_detection H5 video live detection product is composed of two interfaces, which is mainly used in the H5 scene to record and upload a video for the detection of the living body. Compared to the two ways that APP has action check + silent picture living and silent picture living, the H5 video live scheme is more flexible than the APP scheme, and is more secure than the single picture living style. The main functions are as follows: Speech check code: in order to prevent the user from submitting a non current operation video, a number is randomly assigned when the video is recorded. The user needs to read the number, the sound is stored in the video, and is checked at the subsequent recognition to determine whether it is the session. Video multi frame live detection: video recorded and uploaded will be randomly selected for frame analysis in the cloud, and the final living detection score is obtained