Temporal Image Prediction for Video Motion Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current motion estimation techniques in video processing lack sufficient accuracy for optimal use in advanced applications, such as professional settings and high-definition media, where precise motion compensation is required for tasks like upconversion and noise reduction.
Innovation Solution
The method involves accessing a set of motion estimates for each region of a reference frame, predicting temporal image predictions based on these estimates, and blending them with spatial predictions to generate a temporal or spatio-temporal predictor, which improves motion compensation and image processing accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of substance
If motion estimation is used for video compression, then data reduction is achieved, but motion compensation accuracy deteriorates
Solution Approach 1:
The image is divided into multiple blocks, and motion estimation is performed independently for each block. This segmentation allows the system to use simpler estimation techniques for compression while maintaining acceptable accuracy for each local region, resolving the contradiction between data reduction and overall motion compensation accuracy.
Solution Approach 2:
The patent performs motion estimation not only for compression purposes but also generates motion compensation data as a byproduct. This partial action approach uses the same computational framework to achieve both compression (data reduction) and motion tracking (accuracy), allowing the system to address both requirements simultaneously.
2Productivity
If simple motion maps are used for compression, then data encoding efficiency improves, but motion representation accuracy deteriorates
Solution Approach 1:
The video sequence is divided into frames, which are further divided into blocks. Motion maps are generated for each block independently with simplified data structures optimized for compression. This segmentation enables efficient encoding of motion information while maintaining sufficient accuracy for video processing applications.
Solution Approach 2:
The patent uses variable block sizes and adaptive motion estimation parameters to balance compression efficiency and accuracy. By changing parameters such as block dimensions and search ranges dynamically, the system achieves better encoding efficiency without significantly compromising motion representation quality.
Data Source
AI summary
For frames sequentially functioning as a reference frame in a video sequence frame set, a motion estimate set is accessed. One motion estimate characterizes motion associated with pixels of each region of the reference frame in relation to regions of one frame of the set of frames, which is temporally displaced in time with respect to other frames in the set of frames. An additional motion estimate characterizes motion associated with pixels of each reference frame region in relation to a second frame of the set, which is temporally displaced from the one frame and other frames of the frame set. A temporal image prediction set, corresponding to the first and additional motion estimate, is predicted, based on an alignment of the reference frame regions over the frame set. The temporal image predictions are blended and a temporal predictor is generated over the frame set based on the blending.