HDR Image Encoding via Luminance Chrominance Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image and video compression algorithms, such as JPEG and MPEG, struggle with efficiently encoding and decoding high dynamic range (HDR) and high bit-depth images, leading to visible artifacts and high computational power requirements, especially in high-definition resolutions.
Innovation Solution
The proposed solution involves separating HDR images into chrominance and luminance components, using pixel shaders to reconstruct the original texture, and applying error correction techniques to minimize data loss, allowing for efficient compression and decoding compatible with existing standards like JPEG and MPEG.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of substance
If HDR images are encoded using standard compression algorithms (JPEG, MPEG), then compression ratio is improved, but image quality deteriorates due to visible artifacts
Solution Approach 1:
The patent segments HDR image data into multiple components: base layer (clamped LDR version), enhancement layers (fractional color information at different bit depths), and exponent data. This segmentation allows each component to be processed and compressed separately, preserving critical HDR information while enabling efficient compression. The base layer provides fallback compatibility while enhancement layers recover fine details, thus maintaining image quality during compression.
Solution Approach 2:
The patent transforms HDR image representation by separating mantissa (fractional color) and exponent components, and by converting between different bit depth representations (16-bit, 32-bit, 64-bit floating point). This parameter transformation enables the use of standard compression algorithms on transformed data while preserving the ability to reconstruct high-quality HDR images through inverse transformation.
2Manufacturing precision
If HDR images are encoded with high precision, then image quality is improved, but computational power requirements increase
Solution Approach 1:
The patent implements progressive precision encoding where a base layer provides acceptable quality at lower computational cost, and enhancement layers add incremental precision. This allows systems to operate at partial precision (base layer only) when computational resources are limited, while enabling full precision (all enhancement layers) when resources are abundant, thus adapting computational power requirements to available resources.
Solution Approach 2:
The patent performs preliminary clamping of HDR values to LDR range to create the base layer before generating enhancement layers. This preliminary action creates a computationally simple foundation that can be independently decoded at lower precision, while subsequent enhancement layers add detail progressively, reducing the overall computational burden compared to processing full precision data throughout the entire encoding pipeline.
3Adaptability or versatility
If HDR images are compressed using standard algorithms, then compatibility with existing standards is improved, but artifact generation increases
Solution Approach 1:
The patent introduces an intermediary representation where HDR images are decomposed into base layer (LDR-compatible) and enhancement layers (HDR-specific details). Standard compression algorithms operate on this intermediary structure, processing the base layer with conventional methods while preserving enhancement layers that contain critical HDR information. This intermediary structure acts as a bridge between standard compression and HDR preservation, reducing artifacts by preventing the loss of exponent and fractional color information that would occur with direct compression.
4Manufacturing precision
If HDR image data is stored with full precision, then image quality is improved, but data size increases
Solution Approach 1:
The patent segments full-precision HDR data (96 bits per pixel) into separate components: base layer (24 bits), enhancement layers (variable bit depth), and exponent data (8 bits). This segmentation allows selective compression and discarding of less critical components while preserving essential HDR information. The fractional color data in enhancement layers can be stored at reduced precision (e.g., 4-8 bits instead of 16-32 bits) with minimal perceptual impact, significantly reducing overall data size while maintaining color precision where it matters most.
Data Source
AI summary
A method of processing image data includes generating image data including luminance and chrominance data representing a selected object, separating the luminance and chrominance data, storing the separated luminance and chrominance data in corresponding separate spaces in memory, and separately compressing the stored luminance and chrominance data.