Image Encoding Segmentation for Error Resilience
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image encoding methods lack robustness and flexibility in handling errors during wireless transmission, particularly in maintaining high-quality image transmission while reducing data size, and do not efficiently adapt to different image modalities.
Innovation Solution
The method involves splitting images into independent portions, applying a frequency-based transform, quantizing coefficients, and converting them into binary code, with options for varying block sizes and quantization levels, and using pre-filters and probability models tailored to specific image modalities to enhance resilience and compression efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the image is processed as a single unit using traditional encoding methods, then the encoding process is simpler, but errors during transmission can propagate through the entire image and the processing speed is limited
Solution Approach 1:
The image is divided into multiple independent image portions, where each portion is processed separately through the encoding pipeline. This segmentation ensures that errors during transmission are confined to individual portions rather than propagating across the entire image, thereby improving error resilience while maintaining a manageable encoding structure
2Productivity
If the entire image is encoded using a single thread, then the encoding process is simpler to implement, but the processing speed is slower
Solution Approach 1:
The image is divided into multiple independent portions that can be processed simultaneously using multi-threaded implementations. Each thread handles a separate image portion through the complete encoding pipeline, enabling parallel processing that significantly improves encoding speed while keeping each thread's logic relatively simple
Solution Approach 2:
The encoding system dynamically adapts to different processing requirements by allowing flexible configuration of the number of threads and image portions. This dynamic architecture enables the system to optimize processing speed based on available computational resources while maintaining implementation simplicity
3Adaptability or versatility
If a uniform block size is used for all image portions, then the encoding process is more consistent and simpler, but the flexibility to adapt to different image regions and modalities is reduced
Solution Approach 1:
Different block sizes can be assigned to different image portions based on their specific characteristics and requirements. This local adaptation allows the encoding process to optimize for regions with sharp edges using smaller blocks while using larger blocks for uniform regions, improving overall encoding efficiency without requiring complex per-block configuration
Solution Approach 2:
The encoding system provides a universal framework that can handle both uniform and non-uniform block size requirements through a single configurable interface. The same encoding pipeline can process image portions with different block sizes, making the system versatile across different image modalities and applications without requiring separate encoding configurations
4Manufacturing precision
If a single quantisation level is used for the entire image, then the encoding process is simpler and more consistent, but the ability to capture high resolution in important regions while maintaining compression in other regions is lost
Solution Approach 1:
Different quantisation levels can be applied to different image portions based on their importance and characteristics. Regions of interest can be encoded with finer quantisation levels to preserve high resolution details, while less important regions use coarser quantisation levels to achieve higher compression ratios, all within a single configurable encoding process
Data Source
Figure 1a~1B
Figure 2~6
Figure 7~8
AI summary
A method for encoding data defining an image is disclosed. The image is split into a number of image portions. Each portion is segmented into image blocks, each image block in the portion having a uniform block size. A frequency-based transform is applied to each of the image blocks, thereby providing transformed image data in which the image data is represented as coefficients defining a linear combination of predetermined basis functions having different spatial frequencies. The coefficients are quantised, and converted into binary code. Each of the image portions being processed independently of the other image portions.