Staged Motion Vector Search Space Refinement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video codecs face inefficiencies in motion vector estimation, which affects coding efficiency and complexity, particularly in accurately predicting motion between video frames while minimizing signaling overhead.
Innovation Solution
A scheme is introduced to construct a search space for motion vector refinement, involving a first and second search space, where the number and positions of candidate motion vectors in the second search space are determined based on a cost function, allowing for reduced candidate vectors tested while maintaining high coding performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a large number of candidate motion vectors are tested during refinement, then motion vector estimation accuracy is improved, but computational complexity and processing time increase
Solution Approach 1:
The search space is divided into a first search space containing a first plurality of candidate motion vectors and a second search space containing a second plurality of candidate motion vectors. The first search space is determined based on an initial estimate, while the second search space is determined based on candidate motion vectors from the first search space, enabling staged refinement without testing all possible vectors simultaneously
Solution Approach 2:
An initial estimate of the motion vector is obtained first, and a first search space comprising multiple candidate motion vectors is determined based on this initial estimate before performing the final refinement. This preliminary action reduces the search space for the final refinement stage
2Reliability
If motion vector refinement is performed with extensive search spaces, then coding performance is improved, but signaling overhead increases
Solution Approach 1:
The patent extracts and signals only the necessary information (first and second search spaces or candidate motion vectors) rather than signaling all possible candidate motion vectors. This selective extraction maintains coding performance while reducing the amount of information that needs to be transmitted
Solution Approach 2:
Instead of signaling all possible candidate motion vectors (excessive action), the patent signals only the first and second search spaces which contain the necessary candidate vectors for refinement (partial action). This partial signaling is sufficient to achieve the desired coding performance
Data Source
AI summary
The embodiments disclosed herein relate to the construction of a search space for determining a motion vector for a current block of a picture in a video sequence. The search space construction is split into in two main stages, wherein a first and a second partial search space are respectively determined. Based on an initial estimate of a motion vector, a first search space is first constructed. A first and a second candidate motion of the first search space are identified according to a cost function. Based on the first and the second candidate motion vectors, a second search space is constructed. The motion vector for the current block is selected from the candidate motion vectors of the first search space and the second search space.