Motion Vector Precision Transformation for HEVC Storage Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High efficiency video coding (HEVC) systems require increased storage capacity for bidirectional prediction due to the need to retain reference images with fractional pixel accuracy, leading to higher cache storage demands compared to unidirectional prediction.
Innovation Solution
An encoding device that transforms motion vectors with fractional pixel accuracy into integer pixel accuracy for bidirectional prediction, forming predicted images using these transformed vectors and encoding them using a recursively divided block structure, while optionally limiting prediction to either forward or backward prediction in largest coding units.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If motion vectors with fractional pixel accuracy are used for bidirectional prediction, then prediction accuracy is improved, but storage capacity of cache is increased
Solution Approach 1:
The patent changes the precision parameter of motion vectors from fractional pixel accuracy to integer pixel accuracy. This parameter change reduces the storage capacity required in the cache while maintaining acceptable prediction accuracy through compensation techniques.
Solution Approach 2:
The patent introduces an intermediary process where motion vectors with integer pixel accuracy are used as a simplified representation, and compensation is applied to achieve the effect of higher precision without requiring the full storage overhead of fractional pixel accuracy vectors.
2Measurement precision
If reference images in two directions are retained in cache for bidirectional prediction, then prediction quality is improved, but storage capacity is increased
Solution Approach 1:
The patent changes the precision parameter of motion vectors from fractional pixel accuracy to integer pixel accuracy. This parameter change reduces the storage capacity required in the cache while maintaining acceptable prediction accuracy through compensation techniques.
Solution Approach 2:
The patent introduces an intermediary process where motion vectors with integer pixel accuracy are used as a simplified representation, and compensation is applied to achieve the effect of higher precision without requiring the full storage overhead of fractional pixel accuracy vectors.
3Quantity of substance
If motion vectors with fractional pixel accuracy are transformed into integer pixel accuracy, then storage capacity is reduced, but processing complexity is increased
Solution Approach 1:
The patent applies preliminary action by transforming motion vectors to integer pixel accuracy before the motion compensation process. This preprocessing step simplifies subsequent operations and reduces storage requirements, with the transformation complexity being manageable compared to the overall system benefits.
Data Source
AI summary
The present disclosure relates to an encoding device and an encoding method capable of reducing a storage capacity necessary for bidirectional prediction. An accuracy transform unit transforms a motion vector with fractional pixel accuracy of an input image into a motion vector with integer pixel accuracy of the image for each of prediction directions for bidirectional prediction of the input image. An average unit forms a predicted image of the input image for bidirectional prediction on the basis of the motion vector with integer pixel accuracy transformed by the accuracy transform unit for each prediction direction. The present disclosure is applicable to an encoding device of high efficiency video coding (HEVC) system, for example.