Depth Generation Using Motion Vector Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for generating depth information in 3D displays often result in errors, particularly when the foreground is static while the background moves, leading to reverse depth issues and increased memory burden due to long delayed frames.
Innovation Solution
A depth generation method that calculates local and global motion vectors in a 2D to 3D image conversion device, using a peripheral display region to determine the global motion vector while excluding the central region, thereby enhancing accuracy and reducing memory load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional depth generation methods are used to calculate depth information from motion objects, then depth information can be obtained, but reverse depth errors occur when foreground is static and background moves
Solution Approach 1:
The image frame is divided into multiple blocks, and each block is processed independently to calculate local motion vectors. This segmentation allows the system to handle different motion characteristics in different regions, preventing reverse depth errors by treating foreground and background motions separately rather than as a single global motion.
Solution Approach 2:
Different processing approaches are applied to different regions: local motion vectors are calculated for each block to capture local motion details, while a global motion vector is calculated from peripheral blocks to represent overall scene motion. This local quality approach ensures that foreground objects maintain their correct depth relationship while background motion is properly compensated.
2Measurement precision
If delayed frames are used to generate depth information according to different times that delayed frames enter human eyes, then depth information can be generated, but memory burden is excessively increased when delayed period is relatively long
Solution Approach 1:
The invention extracts only the essential motion information (motion vectors) from the image frames rather than storing entire delayed frames. By calculating local and global motion vectors from current and reference frames, the system obtains sufficient depth information while minimizing memory requirements, avoiding the need to buffer multiple complete delayed frames.
Data Source
AI summary
A depth generation method adapted for a 2D to 3D image conversion device is provided. The depth generation method includes the following steps. Motion vectors in an image frame are obtained by motion estimation. A global motion vector of the image frame is obtained. Motion differences between the motion vectors of each block and the global motion vector are calculated. A depth-from-motions of each block is obtained based on the motion differences. Furthermore, a depth generation apparatus using the same is also provided.