Pixel Prediction Method Using Integer Motion Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video compression systems face high complexity and increased decoding complexity when using multiple frames for prediction, particularly in H.264 standards, which complicates the interpolation process and increases computational demands.
Innovation Solution
A method is introduced to calculate motion vectors from downscaled reference vectors, allowing for prediction of a B-frame block from two P-frames with reduced complexity by rounding one motion vector to integer pixel positions, thereby eliminating the need for interpolation in at least one frame and compensating with the other vector, resulting in efficient coding performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple frames are used for prediction in H.264 standards, then prediction accuracy is improved, but decoding complexity increases
Solution Approach 1:
The patent extracts the interpolation operation from one of the prediction paths by rounding the motion vector to integer pixel positions, eliminating the need for complex fractional pixel interpolation in that frame while maintaining prediction accuracy through the other frame's interpolation
Solution Approach 2:
The patent changes the parameter precision of motion vectors by rounding from fractional pixel values to integer pixel values for at least one frame, thereby simplifying the interpolation calculation while compensating through corrected motion vectors to maintain overall prediction accuracy
2Measurement precision
If interpolation is performed for fractional pixel positions, then prediction precision is improved, but computational demand increases
Solution Approach 1:
The patent removes the interpolation operation from one prediction path by using rounded integer motion vectors, eliminating unnecessary computational operations while preserving prediction precision through the other frame's interpolation and error compensation
Solution Approach 2:
The patent uses a simplified integer-based motion vector approximation instead of precise fractional pixel interpolation for at least one frame, accepting a small loss in precision that is compensated by the other frame's accurate interpolation, thereby reducing computational cost
Data Source
AI summary
A method including: calculating a prediction of a first block of pixels in a second video frame with a temporal position between a first and a third video frame from a first motion vector of pixel position components relative to spatial pixel positions of the first block referring the first video frame and a third motion vector of pixel position components relative to spatial pixel positions of the first block referring the third video frame, wherein a second block of pixels in the third video frame spatially collocated with the first block is predicted from a second motion vector of pixel position components relative to spatial pixel positions of the second block referring the first video frame.