Edge Denoising via Projected Pixel Gaussian Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for denoising digital images often introduce artifacts and blurring, and existing demosaicing techniques suffer from issues like the zipper effect and color artifacts, necessitating an alternative approach that effectively mitigates these shortcomings.
Innovation Solution
A method and system that employs a hardware processor to determine the direction of an edge in a raw image, project neighboring pixels onto this edge, apply weighted intensities using a 1D Gaussian filter, and adjust pixel values to denoise the image, while also addressing demosaic artifacts through a cross bilateral filter with an improved guide.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If conventional denoising techniques are applied to reduce noise in digital images, then noise levels decrease, but artifacts and blurring are introduced
Solution Approach 1:
The patent applies different processing strategies to different regions of the image based on local characteristics. Edge pixels are identified and processed differently from non-edge pixels, with edge-preserving filtering applied selectively to maintain sharp transitions while reducing noise in uniform regions.
Solution Approach 2:
The image processing is segmented into distinct stages: edge detection, edge classification, selective filtering, and artifact removal. Each stage processes specific aspects of the image independently, allowing targeted noise reduction without affecting overall image quality.
2Loss of information
If demosaicing process is applied to convert raw images to viewable format, then color information is reconstructed, but zipper effect and color artifacts are introduced
Solution Approach 1:
The patent performs preliminary denoising and edge detection on the raw image data before the demosaicing process. By preparing the data in advance and identifying edge structures beforehand, the subsequent demosaicing algorithm can interpolate color values more accurately without creating zipper effects or color artifacts.
Solution Approach 2:
The patent introduces an intermediate processing stage that bridges raw image data and final demosaiced output. This intermediate stage includes edge-aware filtering and artifact suppression that mediates between the incomplete color samples from the CFA and the final full-color image, preventing artifact propagation.
Data Source
AI summary
Method and system for denoising an edge in a raw digital image are described. First, a direction of a normal to the edge near a pixel p is determined, and neighboring pixels are projected to the normal to the edge, forming projected pixels. Then weighted intensities of the neighboring pixels are determined, including set of weights. One dimensional Gaussian filter centered on the pixel p and acting on the projected pixels is applied, and intensities of the neighboring pixels are adjusted according to the set of weights, resulting in a denoised value z of the pixel p using the weighted intensities, thereby denoising the edge of the raw image expediently.