Pixel Value Up-sampling for HDR Video Banding Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High Dynamic Range (HDR) video compression technologies face challenges in removing color banding artifacts due to insufficient bitdepth, which affects image quality and requires additional processing to achieve higher bitdepth without introducing banding artifacts.
Innovation Solution
A method for up-sampling pixel values from a lower bitdepth to a higher bitdepth by identifying groups of adjacent pixels with equal values and interpolating new values based on edge pixel values, allowing for controlled adjustment of pixel values to create diversity and remove banding effects, while also encoding messages to manage banding artifacts during transmission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If pixel values are converted from lower bitdepth to higher bitdepth, then image quality is improved, but color banding artifacts are introduced
Solution Approach 1:
The patent applies preliminary action by identifying groups of adjacent pixels with equal values before the bitdepth conversion is complete, and pre-calculating interpolated values based on edge pixels. This preparation step ensures that when pixel values are adjusted to higher bitdepth, the transitions are smooth and free from banding artifacts.
Solution Approach 2:
The patent applies local quality by treating different regions of the image differently - specifically, it identifies and processes groups of adjacent pixels with equal values using local interpolation based on neighboring edge pixels. This localized approach ensures that banding is removed in gradient areas while preserving sharp edges and details elsewhere in the image.
2Measurement precision
If bitdepth is increased to remove banding artifacts, then image quality improves, but processing complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the image processing into distinct stages: first identifying groups of adjacent pixels with equal values, then selecting appropriate edge pixels, and finally interpolating values only for specific pixel groups. This segmented approach reduces processing complexity by focusing computational effort only where banding artifacts are likely to occur, rather than processing the entire image uniformly.
Solution Approach 2:
The patent applies self-service by using the image's own edge pixel values to generate interpolation data for removing banding artifacts. The edge pixels within the image itself provide the reference information needed to reconstruct smooth gradients, eliminating the need for external reference data or complex algorithms.
3Object-generated harmful factors
If pixel values are adjusted to create diversity, then banding artifacts are removed, but image processing time increases
Solution Approach 1:
The patent applies preliminary action by pre-identifying groups of adjacent pixels with equal values and pre-selecting edge pixels before the actual interpolation process. This preliminary organization of data reduces processing time during the main conversion operation, as the computational work is structured and ready to execute efficiently.
Solution Approach 2:
The patent applies partial action by focusing interpolation only on groups of adjacent pixels with equal values where banding is most likely to occur, rather than processing every pixel in the image. This selective approach removes banding artifacts effectively while minimizing processing time by avoiding unnecessary computations in areas that don't require intervention.
Data Source
AI summary
A method for managing a picture including pixels. In one aspect, a receiving device converts pixel values of the pixels represented with a first bitdepth into pixel values represented with a second bitdepth where the first bitdepth is smaller than the second bitdepth. The receiving device identifies a first group of pixels including a first pixel and a second pixel. The receiving device identifies a first and second edge pixel, derives a first and second pixel values based on the edge values of the first and second edge pixels, and estimates a first estimated pixel value for the first pixel based on the derived first pixel and second pixel value.