2D to 3D Image Conversion via Segmented Depth Map Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current 2D to 3D video conversion technologies face high computational complexity and unsatisfactory image quality, making them unsuitable for long-term usage.
Innovation Solution
A method that analyzes 2D image content for predefined indicators to generate depth maps, combining them into a single depth map for generating 3D image content, which reduces computational complexity and enhances image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If motion vectors and semantic video content analysis are used to reconstruct 3D objects, then image quality is improved, but computational complexity increases significantly
Solution Approach 1:
The patent segments the depth map generation process into multiple independent indicator analyses (color-based segmentation, linear perspective detection, motion vector analysis, semantic content analysis). Each indicator is processed separately to generate individual depth maps, which are then combined. This segmentation allows the system to selectively apply computational methods based on scene requirements, reducing overall computational complexity while maintaining image quality.
Solution Approach 2:
The patent implements partial action by not requiring all four indicators to be computed for every frame. Instead, the system determines which indicators are most relevant for each specific scene and computes only those, generating depth maps from a subset of indicators. This selective approach significantly reduces computational complexity while maintaining satisfactory image quality by focusing computational resources on the most impactful depth cues.
2Measurement precision
If multiple depth maps are generated and combined, then image quality is improved, but device complexity increases
Solution Approach 1:
The patent divides the complex depth map generation task into four separate indicator-based processing streams (color, perspective, motion, semantics). Each stream independently generates a depth map component, which is then combined through weighted averaging. This segmentation simplifies the overall processing architecture by creating modular, independent processing paths that can be implemented efficiently.
Solution Approach 2:
The patent merges multiple depth maps generated from different indicators into a single combined depth map using weighted averaging. This merging process integrates information from color-based segmentation, linear perspective, motion vectors, and semantic content analysis into a unified depth representation. The combination approach leverages the strengths of each indicator while distributing computational load, improving image quality without proportionally increasing device complexity.
3Measurement precision
If frame to frame movement analysis is performed, then 3D reconstruction accuracy is improved, but processing time increases
Solution Approach 1:
The patent applies partial action by selectively computing motion vectors only when frame-to-frame movement analysis is beneficial for the current scene. The system evaluates scene characteristics and determines whether motion-based depth estimation is appropriate, computing motion vectors only for relevant frames rather than processing every frame uniformly. This approach maintains 3D reconstruction accuracy for dynamic scenes while reducing processing time for static or less motion-critical content.
Data Source
AI summary
A method to convert two-dimensional (“2D”) image content into three-dimensional (“3D”) image content for display on a display device, comprises the steps of: analyzing the 2D image content for predefined indicators and generating a depth map for each of the predefined indicators; determining a combined depth map as a function of the generated depth maps; and generating the 3D image content for display on the display device as a function of the combined depth map.