Depth Map Intra Coding via Discontinuity Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional depth map compression methods, particularly in intra coding, are inefficient due to the piecewise smooth nature of depth maps, leading to increased bit usage and artifacts at depth discontinuities, which degrades coding quality.
Innovation Solution
The introduction of Depth Discontinuity-based Representation (DDR) and Interpolation-based Intra Prediction (IIR) methods, which represent depth maps as smooth partitions with discontinuities and use bilinear functions for efficient compression and reconstruction, reducing arithmetic complexity and error propagation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If conventional DCT-based intra coding methods are used for depth map compression, then the coding process is simple and straightforward, but the method requires a considerable number of bits to cope with depth discontinuities and generates artifacts at discontinuities, degrading coding quality
Solution Approach 1:
The patent divides the depth map into multiple partitions based on detected depth discontinuities. Each partition is then coded separately using DCT, allowing the smooth regions to be efficiently compressed while preserving discontinuities at partition boundaries. This segmentation resolves the contradiction by maintaining simple DCT-based coding within each partition while improving overall coding quality through adaptive partitioning.
Solution Approach 2:
The patent applies different coding strategies to different regions of the depth map based on local characteristics. Smooth regions are coded using standard DCT with fewer bits, while regions containing depth discontinuities are identified and handled specially with increased bit allocation or different transformation methods. This local quality approach resolves the contradiction by optimizing coding quality where needed while maintaining simplicity in smooth regions.
2Ease of manufacture
If conventional DCT-based intra coding methods are used for depth map compression, then the implementation is straightforward, but the method uses a considerable number of bits to represent depth discontinuities, increasing bit rate
Solution Approach 1:
By segmenting the depth map into partitions based on discontinuity detection, the patent reduces the total bit rate. Smooth partitions can be represented with fewer bits since they lack complex discontinuities, while only the necessary discontinuity information at partition boundaries is encoded. This maintains implementation simplicity while reducing overall bit rate compared to uniform DCT coding of the entire depth map.
Solution Approach 2:
The patent extracts and separately codes the discontinuity information from the depth map. By identifying and isolating discontinuity locations and characteristics, the method removes the need to represent all discontinuities using full DCT coefficients, thereby reducing the number of bits required while maintaining implementation feasibility through selective coding.
3Productivity
If conventional DCT-based intra coding methods are used for depth map compression, then the coding process is efficient, but the method produces artifacts at depth discontinuities, degrading reconstruction quality
Solution Approach 1:
The patent segments the depth map into partitions aligned with depth discontinuities. By ensuring that partition boundaries coincide with actual depth edges, the method prevents DCT artifacts from appearing at discontinuities, as each partition's DCT transformation operates on homogeneous regions. This maintains coding efficiency through fast DCT while dramatically improving reconstruction quality by eliminating artifacts.
Solution Approach 2:
The patent applies quality-enhancing techniques locally at discontinuity regions. Where depth discontinuities are detected, the method adjusts coding parameters, increases bit allocation, or applies alternative transformation methods specifically at these locations, while maintaining efficient standard DCT coding in smooth regions. This local quality approach eliminates artifacts at discontinuities without sacrificing overall coding efficiency.
Data Source
AI summary
A method for encoding depth map image involves dividing the image into blocks. These blocks are then classified into smooth blocks without large depth discontinuities and discontinuous blocks with large depth discontinuities. In the discontinuous blocks, depth discontinuities are represented by line segments and partitions. Interpolation-based intra prediction is used to approximate and compress the depth values in the smooth blocks and partitions. Further compression can be achieved with of depth-aware quantization, adaptive de-blocking filtering, scale adaptive block size, and resolution decimation schemes.