Pixel Group Quantization for HDR Image Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for encoding high dynamic range (HDR) images result in significant information loss due to uniform quantization of all pixel values in the residual image, leading to coarse quantization and artifacts like blockiness and banding, especially when only a few pixels have high or low values.
Innovation Solution
Grouping pixels in the residual image into multiple pixel groups based on their characteristics and applying pixel group-specific quantization parameters, allowing for more precise mapping of pixel values within each group, thereby reducing information loss and improving encoding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If uniform quantization is applied to all pixel values in the residual image, then the encoding process is simple and compatible with legacy encoders, but significant information loss occurs and artifacts like blockiness and banding appear
Solution Approach 1:
The patent divides the residual image into multiple pixel groups based on intensity ranges (e.g., dark regions, mid-tone regions, bright regions). Each pixel group is then quantized separately using group-specific quantization parameters, allowing fine quantization for most pixels while using coarser quantization only where necessary, thus reducing overall information loss while maintaining encoding simplicity.
Solution Approach 2:
The patent applies different quantization characteristics to different pixel groups based on their local intensity characteristics. Dark region pixels, mid-tone pixels, and bright region pixels each receive quantization tailored to their specific range, preserving important visual information in each region while maintaining overall encoding efficiency and compatibility.
2Productivity
If coarse quantization is used to reduce bandwidth consumption, then encoding efficiency improves, but image quality deteriorates with visible artifacts
Solution Approach 1:
By segmenting pixels into intensity-based groups, the patent enables selective quantization precision. Most pixels in typical regions (mid-tones and dark regions) can use coarser quantization for efficiency, while pixels in critical regions (bright highlights, transition zones) receive finer quantization, optimizing the balance between encoding efficiency and reconstruction accuracy.
Solution Approach 2:
The patent changes the quantization parameters dynamically based on pixel intensity groups. Different quantization step sizes and mapping functions are applied to different groups, allowing the system to adapt quantization precision to local image characteristics, thereby maintaining high encoding efficiency while minimizing visible artifacts in the reconstructed image.
3Loss of information
If pixel group-specific quantization parameters are used, then information loss is reduced and image quality improves, but the complexity of the encoding process increases
Solution Approach 1:
The patent organizes pixels into a small number of intensity-based groups (e.g., dark, mid-tone, bright regions), which limits the number of different quantization parameter sets that must be managed. This segmentation approach preserves image quality through group-specific quantization while keeping the overall process complexity manageable through the use of a limited number of quantization profiles.
4Manufacturing precision
If fine quantization is applied to preserve image quality, then reconstruction accuracy improves, but bandwidth consumption increases
Solution Approach 1:
The patent applies fine quantization only to specific pixel groups that require high precision (such as bright region pixels and pixels at intensity transitions), while applying coarser quantization to large areas of uniform or low-importance regions. This selective approach maintains reconstruction accuracy where it matters most while minimizing overall bandwidth consumption.
Solution Approach 2:
Different quantization precisions are applied locally to different pixel groups based on their intensity characteristics and visual importance. This allows the system to allocate bandwidth efficiently, providing fine quantization to critical regions that contribute most to perceived image quality while using coarser quantization in less critical areas, thereby optimizing the trade-off between reconstruction accuracy and bandwidth usage.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present invention relates generally to images. More particularly, an embodiment of the present invention relates to the pixel group segmented quantization and de-quantization of the residual signal in layered coding of high dynamic range images. By assigning the pixels in the residual image to different pixel groups based on the pixel value of the corresponding pixel in the decoded base layer signal, and by applying pixel group quantizing functions to assigned pixels a more efficient coding can be achieved.