Depth Map Refinement via Object Segmentation and Edge Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current 2D-to-3D conversion methods face challenges with low-resolution and noisy depth maps, which affect the quality and alignment of depth information for depth image-based rendering, particularly due to rough pixel mapping and inaccurate depth values around objects.
Innovation Solution
A method and system for refining depth maps by identifying objects through edge detection and assigning depth values based on intensity gradients and connectivity analysis, using a segmentation module and processing module to enhance resolution and accuracy, and applying bilateral filtering or connectivity-based filtering to improve boundary accuracy and reduce noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 2D-to-3D conversion is performed using traditional methods, then depth information can be generated, but the depth map resolution remains low and noisy
Solution Approach 1:
The patent segments the image into multiple objects using edge detection and connectivity analysis. By dividing the image into distinct object regions, the system can assign depth values to each object independently, improving both resolution and accuracy of the depth map while reducing noise through region-based processing.
Solution Approach 2:
The patent transitions from 2D image processing to 3D depth map generation by introducing a depth dimension. It uses edge detection in the 2D domain to inform 3D depth assignment, and employs bilateral filtering that considers both spatial and intensity dimensions to refine the depth map, effectively adding depth information while maintaining accuracy.
2Measurement precision
If edge detection is performed to identify objects, then object boundaries can be detected, but noise and false edges are introduced
Solution Approach 1:
The patent implements feedback through iterative refinement processes. The bilateral filtering step uses the detected edges as input and refines them by considering intensity gradients and spatial connectivity. The connectivity analysis provides feedback to validate edges, retaining only those that form continuous boundaries, thereby reducing false edges while maintaining detection accuracy.
Solution Approach 2:
The patent changes parameters by using adaptive thresholding in bilateral filtering and adjusting connectivity criteria based on intensity gradients. By dynamically modifying detection parameters based on local image characteristics, the system improves edge detection accuracy while suppressing noise and false edges through parameter optimization.
3Loss of information
If depth values are assigned to all pixels, then complete depth information is obtained, but computational complexity increases
Solution Approach 1:
The patent segments the image into objects first, then assigns depth values to each object region. This segmentation approach allows the system to process depth information at the object level rather than individual pixel level, reducing computational complexity while maintaining complete depth information coverage through region-based depth assignment.
Solution Approach 2:
The patent uses partial action by focusing computational resources on object boundaries and regions of interest identified through edge detection. Rather than uniformly processing all pixels, the system applies intensive processing only where needed (at edges and object boundaries) while using connectivity analysis to efficiently fill in depth information for remaining regions, reducing overall complexity.
Data Source
AI summary
A method and a system for generating depth information associated with an image. The system comprises a segmentation module arranged to identify a plurality of objects represented by the image, wherein each object adjacent to another object are identified by an edge arranged to substantially separate the object and the adjacent object; and a processing module arranged to generate the depth information by assigning a depth value for each of the plurality of objects, wherein the depth value is arranged to represent a distance between the object and a reference point in a three-dimensional space.