Region-Based Global Motion Models for Video Encoding and Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression technologies face inefficiencies in communicating global motion information to decoders, leading to suboptimal compression efficiency.
Innovation Solution
The implementation of global motion vectors in picture headers, such as the Picture Parameter Set (PPS) and Sequence Parameter Set (SPS), allows for efficient signaling of global motion models, reducing the need for motion vector information within blocks and improving prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If motion vector information is included within each block, then prediction accuracy is improved, but the amount of data required increases and compression efficiency decreases
Solution Approach 1:
The patent merges motion vector information at the picture level rather than block level. By including a single motion vector in the picture header that applies to the entire picture, the system combines the prediction function across all blocks, eliminating redundant motion vector data while maintaining prediction accuracy for images with uniform motion characteristics.
Solution Approach 2:
The picture-level motion vector serves multiple blocks simultaneously, making it a universal parameter that can be applied across the entire picture. This multi-functional approach allows one motion vector to perform the prediction function for all blocks, reducing data requirements while maintaining predictive capability.
2Productivity
If global motion model is signaled in picture header, then compression efficiency is improved, but device complexity increases
Solution Approach 1:
The motion vector is signaled in advance in the picture header before block processing begins. This preliminary signaling allows the decoder to prepare the motion compensation parameters ahead of time, simplifying the subsequent block processing steps and improving overall compression efficiency without adding significant complexity to the decoding operation.
Data Source
AI summary
A decoder includes circuitry configured to receive a bitstream, extract a header, determine, using the header, a global motion model, and decode a current block of a current frame using the global motion model. Related apparatus, systems, techniques and articles are also described.