3D HDR Image Encoding via Tapestry Representation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image processing technologies face challenges in effectively encoding and decoding 3D high-dynamic range images, particularly in handling occluded regions and maintaining image quality across multiple views, which leads to artifacts and inefficient data transmission in autostereoscopic displays.
Innovation Solution
The method employs a tapestry representation that combines multiple views and displacement maps to create a single image with occluded regions included, using inverse displacement and interpolation operations to reconstruct views, and encodes this data efficiently for backward compatibility with standard dynamic range systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional image processing technologies are used to encode and decode 3D HDR images, then data transmission can be performed, but artifacts are generated and image quality deteriorates in occluded regions
Solution Approach 1:
The patent segments the 3D HDR image data into multiple views and represents them using a tapestry structure combined with displacement maps. This segmentation allows each view to be processed and transmitted separately, reducing artifacts in occluded regions by avoiding the blending issues that occur in conventional approaches. The displacement maps encode positional information for each pixel, enabling accurate reconstruction without cross-contamination from occluded areas.
Solution Approach 2:
The patent introduces a new dimensional representation by combining traditional multi-view imagery with displacement map data. This creates a tapestry representation that adds a spatial dimension to the encoding process, allowing pixels to be precisely positioned in 3D space. This dimensional enhancement enables the system to distinguish between occluded and visible regions, thereby reducing artifacts while maintaining image quality.
2Reliability
If multiple views are processed separately to maintain image quality, then data transmission efficiency decreases
Solution Approach 1:
The patent merges multiple views into a single tapestry representation, combining the advantages of separate view processing with the efficiency of unified data transmission. The displacement maps provide the necessary spatial information to reconstruct individual views from the merged tapestry, allowing efficient transmission of a single data structure that contains all view information without sacrificing image quality.
Solution Approach 2:
The tapestry representation serves multiple functions simultaneously: it stores all view information, encodes spatial relationships through displacement maps, and enables both efficient transmission and high-quality reconstruction. This multi-functional approach eliminates the need for separate processing of each view while maintaining the ability to reconstruct high-quality images, thereby improving both transmission efficiency and image quality.
3Loss of information
If occluded regions are filled with pixels from other views, then image completeness improves, but artifacts are generated
Solution Approach 1:
The patent performs preliminary encoding of displacement maps that capture the exact positional information of pixels before reconstruction. This preliminary action allows the decoder to accurately place pixels in their correct positions without needing to fill occluded regions with potentially incorrect data from other views. The displacement maps provide advance guidance for pixel placement, ensuring completeness while avoiding artifacts.
Solution Approach 2:
The displacement maps act as an intermediary between the tapestry representation and the final reconstructed views. Rather than directly copying pixels from one view to fill occluded regions, the displacement maps mediate the process by providing precise positional information. This intermediary layer ensures that pixels are placed correctly, maintaining image completeness while preventing the artifacts that would result from improper pixel filling.
Data Source
AI summary
Representation and coding of multi-view images using tapestry encoding are described for standard and enhanced dynamic ranges compatibility. A tapestry comprises information on a tapestry image, a left-shift displacement map and a right-shift displacement map. Perspective images of a scene can be generated from the tapestry and the displacement maps. Different methods for achieving compatibility are described.