Stereo Vision Depth Map Segmentation for Computational Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for generating depth maps face precision issues and require significant computing resources, often necessitating specialized hardware, which results in poor image resolution and difficulty in identifying precise object locations, limiting their reliability for further processing.
Innovation Solution
A method and apparatus that stabilize segmentation and depth calculations by defining various segment types and using criteria for efficient and accurate depth map updating, focusing on computing changes between consecutive frames rather than recomputing entire depth maps, and leveraging GPU computing for parallel processing to enhance accuracy and reduce computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If specialized hardware is used to generate depth maps, then computing resources are sufficient, but device complexity increases and portability decreases
Solution Approach 1:
The patent replaces specialized hardware systems with a software-based depth map generation method that runs on general-purpose processors. The system uses image processing algorithms and computational methods instead of dedicated hardware circuits, thereby reducing device complexity while maintaining depth map generation capability through software implementation on standard computing platforms.
2Measurement precision
If entire depth maps are recomputed for each frame, then depth map accuracy is maintained, but computational complexity and processing time increase
Solution Approach 1:
The patent divides the depth map computation into segments by identifying and processing only changed regions between consecutive frames. Instead of recalculating the entire depth map, the system segments the image into static and dynamic regions, computing depth only for changed areas. This segmentation approach maintains accuracy for moving objects while significantly reducing overall computational complexity and processing time.
Solution Approach 2:
The patent applies partial action by computing depth maps only for necessary regions rather than the complete image. The system identifies changed regions and performs depth computation exclusively on those areas, using partial computation instead of full recomputation. This approach maintains sufficient accuracy for the changed portions while reducing unnecessary computational overhead in static regions.
3Productivity
If image segmentation is performed to identify changed regions, then computational efficiency improves, but processing complexity increases
Solution Approach 1:
The patent performs preliminary image segmentation and change detection before depth map computation. By pre-identifying changed regions through segmentation in the first stage, the system prepares the data structure and region masks that will be used in subsequent depth computation. This preliminary action organizes the processing workflow and enables efficient targeted computation, reducing overall complexity despite the added segmentation step.
Data Source
AI summary
A method and apparatus for segmenting an image are provided. The method may include the steps of clustering pixels from one of a plurality of images into one or more segments, determining one or more unstable segments changing by more than a predetermined threshold from a prior of the plurality of images, determining one or more segments transitioning from an unstable to a stable segment, determining depth for one or more of the one or more segments that have changed by more than the predetermined threshold, determining depth for one or more of the one or more transitioning segments, and combining the determined depth for the one or more unstable segments and the one or more transitioning segments with a predetermined depth of all segments changing less than the predetermined threshold from the prior of the plurality of images.