HDR Image Decoding via Baseline and Ratio Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for decoding and post-processing high dynamic range (HDR) images on mobile devices require extensive memory and computational resources due to their limited capabilities in handling images with resolutions higher than typical screen resolutions, especially when performing operations like windowing, scaling, and cropping.
Innovation Solution
The proposed solution involves a decoder that separates the HDR image into a baseline and a ratio image, allowing for partial decoding and post-processing of each component separately, utilizing existing hardware accelerators for standard dynamic range (SDR) image processing, and combining the results to produce a decoded HDR image, which reduces memory and processing requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If HDR images are decoded and post-processed using existing hardware accelerators, then decoding speed and processing efficiency are improved, but memory requirements and computational resource consumption increase
Solution Approach 1:
The HDR image is divided into two separate components: a baseline image and a ratio image. Each component is decoded and post-processed independently using existing hardware accelerators, which reduces the memory footprint and computational resources required compared to processing the full HDR image as a single entity. The baseline image is decoded using standard SDR processing, while the ratio image contains only the necessary HDR information.
Solution Approach 2:
The patent extracts only the essential HDR information from the full image data by separating it into a ratio image that contains luminance ratios and a baseline image. This extraction allows the system to process only the necessary data through hardware accelerators, reducing memory consumption while maintaining decoding speed and HDR capabilities.
2Measurement precision
If HDR images with high resolution are processed on mobile devices, then image quality is improved, but computational complexity and processing time increase
Solution Approach 1:
By segmenting the HDR image into baseline and ratio components, the computational complexity is reduced. The baseline image can be processed using standard, less computationally intensive algorithms, while the ratio image contains compressed HDR information that requires fewer operations to reconstruct the final HDR output.
Solution Approach 2:
The patent applies partial decoding by processing only the necessary portions of the image data through hardware accelerators. Instead of fully decoding the entire high-resolution HDR image, the system decodes the baseline image completely and the ratio image partially, only where needed for the final HDR reconstruction, thereby reducing overall computational complexity.
3Reliability
If full decoding of HDR images is performed, then decoding completeness is improved, but processing time and energy consumption increase
Solution Approach 1:
The patent implements partial decoding by processing the baseline image fully and the ratio image selectively. This partial approach maintains decoding completeness for the essential image information while significantly reducing processing time and energy consumption compared to fully decoding the entire HDR image.
Solution Approach 2:
The baseline image is decoded in advance using standard processing, preparing the foundation for HDR reconstruction. This preliminary action allows the ratio image to be processed more efficiently afterward, as the bulk of the decoding work is already complete, thereby reducing total processing time while maintaining completeness.
Data Source
AI summary
A decoder receives for decoding and post-processing for display an HDR (high dynamic range) image comprising a first coded image (e.g., a JPEG-HDR baseline image) and a second coded image (e.g., a JPEG-HDR ratio image). The first coded image is partially decoded and post-processed according to a post-processing command (e.g., scaling) to output a first decoded and post-processed image. The second coded image is also partially decoded and post-processed according to the post-processing command to output a second decoded and post-processed image. The first and the second decoded and post-processed images are combined to output a decoded HDR image according to the post-processing command.