A Light Field Video Motion Estimation Method Based on Nearest Neighbor Optimization

Abstract

The invention discloses a light field video motion estimation method based on neighbor optimization. The light field video motion estimation method comprises the following steps: S1, respectively calculating neighbor parity blocks of K1 airspaces and K2 time domains relative to a current prediction block of a macro pixel; S2, respectively carrying out availability judgment on K1 spatial domain neighbor parity blocks and K2 time domain neighbor parity blocks according to predefined priorities, and selecting M available neighbor parity blocks as reference blocks; S3, performing correction basedon neighbor macro pixel position constraint on the motion vectors of the M reference blocks to obtain M candidate motion vector predictors; and S4, respectively calculating the costs of the candidatereference blocks derived from the M candidate motion vector predictors and the current coding block by adopting a cost model, and carrying out motion search by taking the candidate motion vector predictor with the minimum cost as the final motion vector predictor to complete motion estimation. According to the method, more code streams can be saved for the whole compression process, and the compression rate of the optical field video is improved.

Description

technical field [0001] The invention relates to the fields of computer vision and digital image processing, in particular to a light field video motion estimation method based on nearest neighbor optimization. Background technique [0002] In recent years, light field video captured by light field cameras has shown great application prospects in computer vision fields such as VR / AR content generation, 6-DoF video, and 3D reconstruction, and has attracted widespread attention. For this reason, MPEG launched the creation of the light field video compression standard. Light field video, each frame of which is a light field image, not only records the light angle and position information, but also records the change of this information with time, with ultra-high resolution and special pixel distribution, which makes relatively Compared with the video collected by traditional equipment, the data volume of light field video has increased dramatically. The surge in data volume ha...

AI Technical Summary

Problems solved by technology

[0003] At present, direct compression methods using inter-frame prediction technology for light field video are still very rare, and the compression efficiency is not high enough; while traditional video compression methods can also compress light field video, but the compression rate is very limited; such as HEVC ( High Efficiency Video Coding)'s own inter-frame prediction method can process light field video to achieve a certain compression effect

However, since the data volume of light field video has increased sharply compared with ordinary video, compressing light field video as traditional video will bring huge computational complexity

Although the full search algorithm using HEVC can provide high compression efficiency, its computational complexity is unacceptable; and the fast search method can greatly reduce the computational complexity, but there is still room for improvement in compression efficiency

In the current method of direct compression of light field video using inter-frame prediction, there is a unified defect: due to the mismatch between the macro-pixel structure in the image and the encoder architecture, the correlation between macro-pixels is not fully utilized, which affects the encoding. efficiency

Images