Distributed Source Coding Prediction Mode Reuse
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional distributed source coding methods for correlated images typically require complex and power-consuming encoders, whereas decoders can tolerate higher complexity, limiting the efficiency of syndrome coding due to the lack of optimal intra and inter prediction mode reuse between the side information and input images.
Innovation Solution
The method employs residual coding for both side information and input images, determining residuals within the same image blocks and reusing coding modes from standardized encoders like H.264/AVC or HEVC, which improves encoding efficiency by leveraging intra and inter prediction, resulting in fewer zero bits and reduced bandwidth usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional DSC methods use standardized encoders for side information, then decoding performance is improved, but encoding complexity and power consumption increase
Solution Approach 1:
The system divides the encoding task into two separate parts: the side information image is encoded using a standardized encoder (H.264/AVC or HEVC) to maintain compatibility and decoding performance, while the input image is encoded using a simplified syndrome encoder. This segmentation allows each component to be optimized independently, resolving the contradiction between decoding performance and encoding complexity.
Solution Approach 2:
The patent introduces prediction modes as an intermediary element that is extracted from the side information encoding process and reused in the input image encoding. This intermediary mechanism allows the system to leverage the computational effort already spent on side information encoding to improve the encoding of the input image, thereby reducing overall encoding complexity while maintaining decoding performance.
2Device complexity
If syndrome coding is used for both images, then encoding complexity is reduced, but prediction mode optimization is lost
Solution Approach 1:
The system performs preliminary encoding of the side information image using a standardized encoder, which automatically determines optimal prediction modes for that image. These pre-determined prediction modes are then reused for encoding the input image, eliminating the need to perform prediction mode optimization again for the input image. This preliminary action resolves the contradiction by maintaining encoding efficiency through reuse while keeping encoding complexity low.
Solution Approach 2:
The side information encoding process serves itself by generating prediction modes that are subsequently reused for the input image encoding. This self-service mechanism allows the system to extract useful information (prediction modes) from the side information encoding and apply it to the input image encoding, thereby improving overall encoding efficiency without requiring additional computational resources for mode optimization.
3Measurement precision
If prediction modes are determined separately for each image, then encoding accuracy is improved, but computational resources are consumed
Solution Approach 1:
The patent makes prediction modes universal by extracting them from the side information encoding and reusing them for the input image encoding. This multi-functionality approach allows the same prediction modes to serve dual purposes: optimizing the encoding of the side information image and subsequently optimizing the encoding of the input image. This resolves the contradiction by maintaining encoding accuracy through prediction mode optimization while reducing computational resource consumption through reuse.
Data Source
AI summary
In a decoder, a desired image is estimated by first retrieving coding modes from an encoded side information image. For each bitplane in the encoded side information image, syndrome bits or parity bits are decoded to obtain an estimated bitplane of quantized transform coefficients of the desired image. A quantization and a transform are applied to a prediction residual obtained using the coding modes, wherein the decoding uses the quantized transform coefficients of the encoded side information image. The estimated bitplanes of quantized transform coefficients of the desired image are combined to produce combined bitplanes. Then, an inverse quantization, an inverse transform and a prediction based on the coding modes are applied to the combined bitplanes to recover the estimate of the desired image.