Edge Shape Enforcement for Depth Image Based Rendering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Depth image based rendering (DIBR) techniques in 3D video streams face challenges in producing high-quality intermediate views due to inaccuracies in depth or disparity information, leading to artifacts and poor visual quality, especially along object edges.
Innovation Solution
A method and device that corrects depth or disparity information by detecting and correcting edges using prior knowledge of edge shape preservation under perspective transformations, applying corrections either before transmission or during the rendering process, and utilizing metrics for evaluating warping errors to enhance image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If depth or disparity information is used for DIBR warping, then intermediate views can be generated, but edge accuracy deteriorates due to depth estimation errors and quantization
Solution Approach 1:
The patent segments the image into edge regions and non-edge regions. Edge detection algorithms identify edge pixels, and separate processing is applied to these regions. The warping process uses corrected depth values specifically for edge pixels, while non-edge regions use standard DIBR warping. This segmentation allows targeted correction of edge accuracy without affecting the overall intermediate view generation capability.
Solution Approach 2:
The patent applies local quality enhancement by detecting edges and applying specific correction operations only to edge regions. The method computes edge error metrics locally and applies depth correction selectively to pixels near detected edges. This local approach improves edge accuracy in the synthesized view without requiring global reprocessing of all pixels, thus resolving the contradiction between generating intermediate views and maintaining edge precision.
2Productivity
If depth quantization is applied to reduce data rate, then transmission efficiency improves, but depth accuracy deteriorates leading to warping errors
Solution Approach 1:
The patent applies preliminary action by performing edge detection and depth correction before the DIBR warping process. The method identifies edge pixels in advance, computes correction terms based on edge error metrics, and adjusts depth values for these pixels before they are used in warping. This preliminary correction prevents warping errors from occurring in the first place, rather than attempting to fix them after warping has been applied.
Solution Approach 2:
The patent implements feedback by computing edge error metrics that compare the warped edge positions with expected positions. These error metrics are used to generate correction terms that are fed back into the depth map. The corrected depth values are then used in subsequent warping operations, creating a feedback loop that continuously improves edge accuracy while maintaining the quantized depth representation for efficient transmission.
3Adaptability or versatility
If standard DIBR warping is applied, then intermediate views are generated, but visual quality deteriorates due to edge distortion and artifacts
Solution Approach 1:
The patent segments the warping process into edge-aware and non-edge portions. Edge detection algorithms identify regions requiring special handling, and these segments are processed with corrected depth values. The segmentation allows the system to maintain standard DIBR functionality for most of the image while applying enhanced processing only where needed, thus improving visual quality without sacrificing intermediate view synthesis capability.
Solution Approach 2:
The patent changes parameters locally in edge regions by computing correction terms that adjust depth values for edge pixels. The method modifies the warping parameters (depth values) specifically for edge regions based on detected edge errors, while leaving non-edge parameters unchanged. This parameter change approach improves visual quality by reducing edge distortion and artifacts in the synthesized intermediate views.
Data Source
AI summary
A method for edge correction of images of a three-dimensional video content, the video content including at least one original view image and at least one depth or disparity map, the method including the following steps: detecting edges in at least one original view image for obtaining original edges; warping the original edges according to the depth or disparity map; detecting a set of warped edges altered by the warping process; and correcting the altered edges for obtaining corrected edges.