A rate-adaptive decision-making method supporting face detection and real-time super-resolution

Abstract

A rate-adaptive decision-making method that supports face detection and real-time super-resolution. Through YCbCr chroma detection and inter-frame cost determination, the face area information and texture motion area information of each macroblock are obtained, and the information of each macroblock is calculated. Interest area level and intra-frame factor; combine the interest area level and intra-frame factor to design interest encoding and quantization scheme, and integrate it into H.264 encoder to encode video at multiple bit rates; cut and encapsulate the encoded video , the client downloads the corresponding video file to the buffer according to the feedback result of the bit rate adaptive decision-making module, completes the super-resolution reconstruction in the buffer, replaces the original video block after re-encoding, and the player plays the video block with enhanced quality . The invention can compress the video size of the encoded human face video image as much as possible while ensuring the subjective quality, so as to achieve the purpose of reducing bandwidth consumption, and at the same time ensure the high bit rate of important areas in the video image, and improve the clarity of the video.

Description

technical field [0001] The invention belongs to the technical field of video transmission, in particular to a code rate adaptive decision-making method supporting face detection and real-time super-resolution. Background technique [0002] With the development of Internet technology and the popularity of smart terminal devices, it is becoming more and more popular to use terminal devices to watch videos online. At present, most video resources take people as the main body. When watching videos online, people will pay more attention to the facial expressions and body movements of the people in the video. Focus on the area, and this small portion of video content determines the quality of the user's viewing experience. With the development of computer vision technology, many artificial intelligence algorithms for face recognition have been proposed, but these algorithms often require powerful computing power support, and the complexity is high, which is not conducive to the r...

AI Technical Summary

Problems solved by technology

[0004] In video transmission and playback, the adaptive bit rate strategy based on fixed rules cannot optimize the performance of the adaptive bit rate transmission system, which has certain limitations and particularities

Images