In-loop Reshaping for HDR Video Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video compression standards are limited in efficiently encoding and decoding high dynamic range (HDR) images, as they often constrain bit depth to 8-10 bits per pixel, which is insufficient for HDR content that spans a wide luminance range from 0.001 to 10,000 cd/m², leading to suboptimal compression efficiency and image quality.
Innovation Solution
The implementation of in-loop adaptive reshaping techniques within video encoders and decoders, which involve generating forward and backward reshaping functions to convert HDR images into a target bit depth, allowing for more efficient compression and decompression while maintaining high image quality, by utilizing perceptual quantization methods that match the human visual system's response to luminance levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If video compression standards constrain bit depth to 8-10 bits per pixel, then device complexity and processing efficiency are improved, but image quality and dynamic range representation deteriorate
Solution Approach 1:
The patent applies parameter changes by transforming the bit depth parameter dynamically. It converts HDR images from high bit depth (10-16 bits) to target bit depth (8-10 bits) using adaptive reshaping functions. The reshaping parameters are adjusted based on local image characteristics such as luminance range and complexity, allowing optimal quality preservation at reduced bit depth while maintaining processing efficiency.
Solution Approach 2:
The patent implements local quality by applying different reshaping strategies to different regions of the image. It divides the HDR image into multiple regions and applies adaptive reshaping functions tailored to each region's characteristics. This ensures that complex regions maintain higher quality while simpler regions use more aggressive compression, resolving the contradiction between overall quality and processing complexity.
2Productivity
If bit depth is reduced to 8-10 bits per pixel, then compression efficiency and processing speed are improved, but the representation of HDR content with luminance range 0.001-10,000 cd/m² deteriorates
Solution Approach 1:
The patent uses parameter changes by dynamically adjusting the reshaping function parameters based on the luminance characteristics of different image regions. The adaptive quantization parameters are optimized to preserve luminance information across the wide HDR range (0.001-10,000 cd/m²) while targeting 8-10 bit depth, thus maintaining compression efficiency without excessive information loss.
Solution Approach 2:
The patent applies dynamics by making the reshaping process adaptive rather than static. The reshaping functions and quantization parameters are dynamically adjusted based on local luminance statistics and image complexity. This dynamic adaptation allows the system to maintain high compression efficiency while preserving critical luminance information across the extensive HDR range.
3Manufacturing precision
If adaptive reshaping functions are implemented, then image quality and HDR representation are improved, but device complexity and processing requirements increase
Solution Approach 1:
The patent applies segmentation by dividing the HDR image into multiple regions and computing separate adaptive reshaping functions for each region. This segmentation allows the complex adaptive processing to be distributed and managed locally, reducing the overall computational burden on the encoder while maintaining high image quality through region-specific optimization.
Solution Approach 2:
The patent implements partial action by applying adaptive reshaping selectively to regions that benefit most from it. Rather than processing the entire image with full adaptive complexity, the system identifies and applies sophisticated reshaping only where needed, using simpler methods elsewhere. This reduces encoder complexity while preserving image quality in critical regions.
Data Source
Figure 1A~1B
Figure 2A
Figure 2B
AI summary
Systems and methods are disclosed for in-loop, region-based, reshaping for the coding of high-dynamic range video. Using a high bit-depth buffer to store input data and previously decoded reference data, forward and backward, in-loop, reshaping functions allow video coding and decoding to be performed at a target bit depth lower than the input bit depth. Methods for the clustering of the reshaping functions to reduce data overhead are also presented.