Adaptive Sparse Filter for Image Debanding in HDR Displays
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies fail to effectively reduce banding artifacts in images, especially in high dynamic range (HDR) content, leading to visual anomalies and poor image quality due to inadequate compression and post-processing techniques.
Innovation Solution
The implementation of adaptive sparse filtering, where the parameters of the sparse filter are adjusted based on the properties of the image bands, allowing the filter to be applied across various orientations and iteratively refined to minimize banding while preserving image edges and details.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If compression techniques are applied to HDR images, then image data can be stored and transmitted efficiently, but banding artifacts appear in the displayed image
Solution Approach 1:
The image processing is divided into separate stages: tone mapping compression followed by debanding filtering. This segmentation allows each stage to optimize for its specific function - compression efficiency and artifact removal respectively - resolving the contradiction between efficient data transmission and artifact-free display
Solution Approach 2:
A debanding filter acts as an intermediary processing step between tone mapping and final display. This mediator removes the banding artifacts introduced by compression while preserving the benefits of efficient data storage and transmission, effectively resolving the harmful side effect
2Object-affected harmful factors
If filtering is applied to remove banding artifacts, then image quality improves, but image edges and details may be blurred
Solution Approach 1:
The debanding filter applies different processing characteristics to different regions of the image - smooth gradient areas receive stronger filtering to remove bands, while edge regions maintain their sharpness. This local differentiation resolves the contradiction by allowing artifact removal without compromising edge precision
Solution Approach 2:
The filter adapts its behavior dynamically based on local image characteristics such as gradient magnitude and orientation. This dynamic adaptation allows the filter to be aggressive where needed (smooth areas with bands) and conservative where precision is critical (edges and details), resolving the quality trade-off
3Productivity
If a fixed filter parameter is used for debanding, then processing is simple and fast, but the filter cannot adapt to different band widths and orientations
Solution Approach 1:
The filter parameters are pre-computed based on local image analysis (gradient orientation, band width estimation) before the actual filtering operation. This preliminary preparation enables the filter to adapt to different conditions without adding significant processing overhead during the main filtering pass, resolving the speed-adaptability contradiction
Data Source
Figure 1~2
Figure 3A
Figure 3B
AI summary
Methods and systems for reducing banding artifacts when displaying images are described. Identified image bands are filtered using an adaptive sparse finite response filter, where the tap-distance in the sparse filter is adapted according to an estimated width of each image band. Image debanding may be performed across multiple pixel orientations, such as rows, columns, a 45-degree angle, or a -45-degree angle. Given a threshold to decide whether sparse filtering needs to be performed or not, an iterative debanding process is also proposed.