Multi-view Image Encoding via View Component Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Multi-view images require significant storage and transmission bandwidth due to redundant data between views, and existing methods struggle to efficiently encode and render these images, especially when generating new views not present in the original data.
Innovation Solution
The method involves separating view-dependent and view-independent components of a multi-view image, using three-dimensional geometry information and texture maps to reconstruct and encode images, allowing for efficient data compression and rendering of original and new views.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multi-view images are stored and transmitted using traditional methods, then complete view information is preserved, but storage space and transmission bandwidth increase significantly
Solution Approach 1:
The patent segments the multi-view image data into two distinct components: view-independent components (common to all views) and view-dependent components (specific to each view). This segmentation allows the system to store and transmit only the essential unique information for each view while reconstructing the complete multi-view dataset through computational processing, thereby reducing overall data volume while preserving view information completeness.
Solution Approach 2:
The patent extracts the view-independent component from the multi-view image data and separates it from the view-dependent components. By taking out the common information that appears across all views and storing it separately, the system eliminates redundant data storage and transmission while maintaining the ability to reconstruct complete view information when needed.
2Manufacturing precision
If the number of captured views increases to generate higher-quality 3D content, then 3D content quality improves, but transmission bandwidth requirements increase
Solution Approach 1:
The patent applies segmentation by dividing the texture information into view-independent and view-dependent components. This allows the system to capture high-quality 3D content from multiple views while transmitting only the segmented components, significantly reducing bandwidth requirements while maintaining the ability to reconstruct high-quality 3D content.
Solution Approach 2:
The patent merges the view-independent component with view-dependent components during the rendering process to reconstruct complete high-quality multi-view images. This combining approach allows the system to transmit less data while reconstructing the full-quality multi-view content, effectively reducing bandwidth requirements without sacrificing 3D content quality.
3Ease of manufacture
If traditional encoding methods are used for multi-view images, then encoding simplicity is maintained, but encoding efficiency and data compression remain insufficient
Solution Approach 1:
The patent implements segmentation of multi-view image data into view-independent and view-dependent components, which enables more efficient encoding by allowing the system to encode the common component once and reuse it across multiple views, rather than encoding each view independently. This approach maintains relative encoding simplicity while significantly improving encoding efficiency and compression ratios.
Data Source
AI summary
The present disclosure provides a multi-view image decoding method including: obtaining, from a bitstream, three-dimensional geometry information indicating a three-dimensional space of a multi-view image, view independent component information indicating a view independent component, which is uniformly applied to every view, and view dependent component information indicating a view dependent component, which is differently applied according to views; determining a view dependent component for a texture map of a current view from the view dependent component information; generating the texture map of the current view from a view independent component of the view independent component information and the determined view dependent component for the current view; and reconstructing a current view image according to a three-dimensional space that is constructed according to the texture map of the current view and the three-dimensional geometry information.