Image Denoising via Pre-computed Patch Distance Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current patch-based image processing methods are computationally intensive and inefficient, particularly for high-resolution images, due to the extensive calculations required for determining distances between patches, which is a limitation for mobile platforms with limited computing resources.
Innovation Solution
The method generates lookup tables and integral images based on comparisons of a digital image to its offsets, allowing for efficient computation of distances between patches, reducing computational complexity and memory footprint by a factor of the number of patches used.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current patch-based processing methods are used to compute distances between patches, then image denoising quality is maintained, but computational complexity and processing time increase significantly
Solution Approach 1:
The patent pre-computes and stores patch distances in a distance map before the actual denoising process. By calculating all pairwise patch distances in advance and storing them in a lookup table, the method eliminates the need for repeated distance computations during denoising operations, significantly reducing real-time computational complexity while maintaining accurate distance measurements for weight calculation
Solution Approach 2:
The patent creates a distance map (copy) that stores pre-computed distance values between patches. Instead of recalculating distances between patches during denoising, the system retrieves pre-stored distance values from this copy, reducing computational complexity from O(N×P×K) to O(1) per pixel processing while preserving measurement precision
2Measurement precision
If current patch-based processing methods are used, then comprehensive patch comparisons are performed, but memory consumption and processing overhead increase
Solution Approach 1:
The patent extracts only the essential distance information from comprehensive patch comparisons and stores it in a compact distance map. By computing and storing only the distance metric (而非 full patch data), the system reduces memory footprint while retaining the necessary information for accurate weight calculation in the denoising process
3Reliability
If extensive patch distance computations are performed, then accurate noise removal is achieved, but power consumption increases on mobile devices
Solution Approach 1:
The patent performs distance computations in advance and stores results in a distance map, eliminating the need for repeated calculations during denoising. This preliminary computation approach reduces the number of processor cycles required during actual image processing, thereby reducing power consumption on battery-powered mobile devices while maintaining noise removal effectiveness
Solution Approach 2:
By creating and using a pre-computed distance map copy, the system avoids repeated expensive distance calculations during denoising operations. The lookup table stores distance values that can be quickly retrieved and used for weight calculation, significantly reducing computational load and power consumption compared to real-time distance computation
Data Source
AI summary
In embodiments of removing noise from an image via efficient patch distance computations, weights are computed for patches of pixels in a digital image, and the computed weights are multiplied by respective offset values of offset images that are pixelwise shifted images of the entire digital image. The weights can be applied to the pixels in the digital image on a patch-by-patch basis to restore values of the pixels. Additionally, the digital image can be pixelwise shifted to generate the offset images of the digital image, and the digital image is compared to the offset images. Lookup tables of pixel values can be generated based on the comparisons of the digital image to the offset images, and integral images generated from the lookup tables. Distances to the patches of pixels in the digital image are computed from the integral images, and the computed weights are based on the computed distances.