Image Processing Apparatus Using Spatial Filters for HDR Contrast
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques fail to effectively address contrast issues between High Dynamic Range (HDR) and Standard Dynamic Range (SDR) images, leading to artifacts and uniform contrast adjustments that do not consider spatial distribution, resulting in suboptimal visual representation.
Innovation Solution
An image processing apparatus that applies spatial filters based on human visual characteristics to generate contrast influence information by filtering brightness, color, and saturation components across different frequency bands, allowing for targeted contrast correction in HDR and SDR images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If wavelet transform is used to generate multi-frequency luminance images, then contrast effects can be addressed, but artifacts occur in the processed image
Solution Approach 1:
The image processing is segmented into multiple frequency bands using band-pass filters, allowing contrast detection to be performed separately for different spatial frequencies. This segmentation enables precise contrast measurement without the artifacts introduced by wavelet transform, as each frequency component is processed independently and then recombined.
Solution Approach 2:
The patent introduces an intermediary processing step using band-pass filters as mediators between the original luminance image and the final contrast-corrected image. These filters selectively extract specific frequency components, enabling accurate contrast detection while avoiding the harmful artifacts that directly applying wavelet transform would introduce.
2Stability of the object's composition
If tone conversion curve adjustment is used to equalize luminance changes, then visual consistency is improved, but spatial distribution of contrast influence is not considered
Solution Approach 1:
The patent applies local quality by calculating contrast influence separately for different spatial regions and frequency bands. Instead of applying a uniform tone conversion curve across the entire image, the system computes contrast influence values for each pixel based on its local neighborhood and frequency characteristics, then applies targeted corrections that preserve spatial variations in contrast effects.
Solution Approach 2:
The patent adds the frequency dimension to the traditional spatial contrast correction approach. By analyzing contrast influence across multiple frequency bands (low, medium, high) in addition to spatial位置, the system achieves more precise control over contrast correction, allowing different correction strategies for different spatial frequencies while maintaining visual consistency.
3Ease of operation
If uniform contrast correction is applied across the entire image, then processing simplicity is maintained, but spatial variations in contrast influence are ignored
Solution Approach 1:
The image is segmented into multiple frequency bands using band-pass filters, allowing contrast detection to be performed separately for different spatial frequencies. This segmentation enables precise contrast measurement without the artifacts introduced by wavelet transform, as each frequency component is processed independently and then recombined.
Data Source
AI summary
An image processing apparatus includes a first filter unit and a first composition unit. The first filter unit generates a plurality of results of filter processing for brightness by individually applying each of a plurality of filters for brightness with different frequency bands to a first brightness component image indicating a distribution of brightness components in a first image. The plurality of filters for brightness are spatial filters based on a visual characteristic of a human related to contrast detection. The first composition unit generates brightness contrast influence information indicating a distribution of degrees of influence of brightness contrast in the first image by compositing together the plurality of results of filter processing for brightness.