Video Compression Motion Vector Estimation Using Depth Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video compression methods are inefficient due to high computational load from iterative motion estimation, often compromising quality by limiting search areas or estimating motion vectors based on neighbor blocks, which affects the efficiency and quality of the compression process.
Innovation Solution
Determining motion vectors based on camera movements using depth maps to project image points into a three-dimensional space, allowing for accurate motion estimation and compensation, including camera movements such as translation, contraction, expansion, and rotation, thereby enhancing the precision of motion vectors and reducing computational requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If iterative motion estimation is used to achieve accurate motion vectors, then measurement precision of motion vectors is improved, but computational load increases significantly
Solution Approach 1:
The patent applies preliminary action by using depth information from the reference frame to pre-calculate expected motion vectors before performing actual motion estimation. The depth-based predicted motion vectors serve as initial guesses that guide the iterative search process, reducing the computational effort needed to achieve accurate motion vectors by starting closer to the optimal solution.
Solution Approach 2:
The patent introduces depth information as an intermediary element that bridges the reference frame and current frame. By using depth maps or depth estimates, the system can predict motion vectors without requiring full iterative search, thus reducing computational load while maintaining accuracy. The depth information acts as a mediator that enables faster motion estimation.
2Productivity
If search area is limited to reduce computational load, then productivity is improved, but measurement precision of motion vectors deteriorates
Solution Approach 1:
The patent uses depth information to pre-determine the likely motion vector direction and magnitude before limiting the search area. This preliminary prediction allows the system to focus the search only in relevant regions, maintaining both speed and accuracy.
Solution Approach 2:
The patent applies local quality by adapting the search strategy based on depth information. Different regions of the image are treated differently according to their depth values, with the search area and precision adjusted locally. This allows efficient compression while maintaining motion vector accuracy in critical regions.
3Productivity
If motion vectors are estimated based on neighbor blocks to reduce computational load, then productivity is improved, but measurement precision of motion vectors deteriorates
Solution Approach 1:
The patent uses depth information to pre-calculate motion vectors for neighbor blocks or to predict motion patterns before processing the current block. This preliminary action provides better initial estimates that improve the accuracy of motion vector estimation for all blocks, not just neighbor blocks.
Solution Approach 2:
The patent makes the depth-based motion prediction mechanism universal by applying it to all blocks in the frame rather than just neighbor blocks. The same depth information processing is used to predict motion vectors throughout the entire image, improving overall compression efficiency while maintaining accuracy.
Data Source
AI summary
A method (800, 820 ), device (900 ) and program product for compression video information is presented. The method comprises the following steps: projecting (803 ) points of a next image to a three-dimensional space using camera parameters and depth map and projecting (805 ) projected points from the three-dimensional space to reference image surface, thereby obtaining motion vectors for estimating changes between the next image and the reference image.