A Key Frame Acquisition Method for Portrait Video System

Abstract

The invention relates to a key frame acquisition method for a portrait video system, comprising the following steps: 1) acquiring video images, and performing key point detection on each frame of images; 2) performing attribute classification on each frame of images according to the detected key points , and calculate the confidence under each attribute; 3) select the key frame of P frame according to the confidence of each frame image under different attributes, as the basis of face recognition. Compared with the prior art, the invention improves the key frame selection process, which not only ensures the speed, but also has higher recognition performance.

Description

technical field

[0001] The invention relates to the technical field of video processing, in particular to a key frame acquisition method for a portrait video system. Background technique

[0002] Face recognition is the core module in the portrait video system. In most portrait systems, the corresponding trajectory will be obtained for passing people. The trajectory of a single person can be represented as a collection of rectangular boxes on K frames of images. The face recognition module in the portrait video system inputs the trajectory of a person and outputs the identity of the person.

[0003] Compared with the recognition module of the portrait image system, the main difference of the portrait video system is that it needs to select the face on the appropriate frame for recognition. The easiest way is to recognize all K frames of images, which will slow down the system speed, because the time overhead of the recognition module is very large (1 frame 1 second, and a...

Images