Last Frame Motion Vector Partitioning for Video Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video compression techniques require significant computational resources and data storage due to the large amount of data in video streams, and existing methods re-partition frames from scratch without utilizing previous frame partitioning, leading to inefficient prediction and increased distortion.
Innovation Solution
The method involves using last frame motion vector partitioning to adjust the partitioning of the current frame based on the motion vectors, allowing for the splitting of blocks into separate regions by shifting prior frame partitions, enabling separate modes and motion vectors for non-square regions and reducing prediction costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If frames are re-partitioned from scratch for each current frame, then prediction accuracy may be improved, but computational resources and processing time increase significantly
Solution Approach 1:
The patent applies preliminary action by determining the partitioning of the current frame based on the already-determined partitioning of the previous frame, rather than performing fresh partitioning from scratch. This reuses previously computed partitioning information and adjusts it using motion vectors, thereby reducing redundant computational operations while maintaining prediction accuracy.
Solution Approach 2:
The patent uses copying by replicating the partitioning structure from the previous frame to the current frame. Instead of recalculating partitions independently, the system copies the prior partitioning configuration and modifies it according to motion compensation, significantly reducing the computational burden of frame partitioning while preserving prediction effectiveness.
2Adaptability or versatility
If the entire frame is re-partitioned for each current frame, then prediction adaptability is improved, but data storage and transmission requirements increase
Solution Approach 1:
The patent determines partitioning information in advance by reusing the previous frame's partitioning structure, which reduces the amount of new partitioning data that needs to be stored and transmitted. This preliminary determination based on historical data minimizes data volume while preserving adaptability through motion-compensated adjustments.
Solution Approach 2:
By copying the partitioning structure from the previous frame and modifying it with motion vectors, the patent avoids generating and storing completely new partitioning data for each frame. This approach maintains prediction adaptability while significantly reducing the quantity of data that must be stored and transmitted compared to full re-partitioning.
3Productivity
If motion vectors are used to shift prior frame partitions, then computational resources are reduced, but prediction precision for complex motions may deteriorate
Solution Approach 1:
The patent applies segmentation by dividing the frame into multiple blocks and applying motion compensation at the block level. Each block can have its own motion vector, allowing the system to handle complex motions more accurately while still benefiting from the computational efficiency of reusing prior partitioning structures. This segmented approach balances computational savings with prediction precision.
Solution Approach 2:
The patent uses local quality by allowing different regions of the frame to have different partitioning characteristics and motion vectors. Instead of applying a uniform partitioning approach, the system adapts the partitioning locally based on motion patterns, thereby maintaining prediction precision in complex motion areas while achieving computational efficiency in simpler regions.
Data Source
AI summary
A first partitioning for partitioning a first region of a frame before a current frame in a video sequence is determined. The first partitioning defines at least two prediction sub-regions of the first region, and the prediction sub-regions include contiguous, non-overlapping groups of pixels. A motion vector that predicts a prediction sub-region of the first region is determined. The first partitioning is modified to a second partitioning by moving at least one border of the first partitioning that is between adjacent prediction sub-regions of the first region by the motion vector. A current region of the current frame that is collocated with the first region is partitioned by the second partitioning, and is encoded or decoded. The technique eliminates the need to re-partition an entire frame each time, and can be used with both rectangular-shaped block partitioning and partitioning using masks and other partitioning techniques.