Region of Interest Video Coding Method and Device Based on Object Detection

Abstract

The invention relates to the technical field of video coding, and discloses a method and device for video coding of a region of interest based on target detection. In the present invention, the video coding method includes the following steps: performing target detection on the original image, generating target information, and obtaining the target area and confidence; converting the target area into an area of ​​interest suitable for encoding; calculating Coding the quantization parameter; performing quantization coding on the region of interest and the region of non-interest according to the corresponding coding quantization parameter. Obtain the region of interest and non-interest region and the corresponding confidence through target detection, and select the corresponding quantization parameters to quantize and encode the region of interest and non-interest region, which can effectively ensure the coding quality and stability of the region of interest. On the premise of reducing the coding quality of the region of interest, the compression code stream of the entire image can be reduced, and the efficiency of video transmission and storage can be improved.

Description

technical field [0001] The present invention relates to the technical field of video coding, in particular to a method and device for video coding a region of interest based on target detection. Background technique [0002] Video surveillance is an important part of the security industry. With the continuous expansion of the video surveillance system and the increase of monitoring points, it is necessary to pay a huge cost for data transmission and storage. Therefore, video information must be efficiently compressed to make full use of network resources, reduce storage space, and achieve cost savings. [0003] In the practical application of video surveillance, people are generally only interested in a certain part of the surveillance scene, thus potentially requiring higher coding quality in the region of interest, while not requiring high coding quality in the region of non-interest. In the previous monitoring system, the entire image was uniformly adjusted for encoding...

AI Technical Summary

Problems solved by technology

There are solutions in the prior art that use H.264’s flexible macroblock sorting to process ROI coding, try to preserve data information for ROI, and minimize coding bits for non-ROI, but this method is only applicable to H. .264 encoding, and additional bit overhead needs to be added, general decoders cannot support flexible macroblock sorting, and require higher access to the decoding end

Images