Multi-Resolution Image Processing for Noise Reduction and Texture Preservation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods struggle to effectively reduce noise in image data without excessively blurring texture regions, as it is difficult to distinguish between noise and actual texture.
Innovation Solution
An electronic device and method that utilizes an image sensor and processor to generate multiple images of varying sizes, calculate parametric images using color and texture pixel values, and synthesize these images to reduce noise while preserving texture details.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If an average filter or Gaussian filter is used to blur the image, then noise is reduced, but texture regions become excessively blurred
Solution Approach 1:
The patent applies different processing strengths to different regions of the image by calculating a texture determination value for each pixel. Regions with high texture determination values (texture regions) receive different treatment compared to regions with low texture determination values (smooth regions), allowing noise reduction in smooth areas while preserving texture details.
Solution Approach 2:
The patent changes the parameter used for noise determination from direct pixel intensity comparison to a texture determination value derived from multiple images at different resolutions. This parameter change enables better discrimination between noise and actual texture, allowing selective noise reduction without excessive blurring.
2Productivity
If image data is synthesized at reduced resolution for regions with small textures, then processing efficiency is improved, but texture regions may be excessively blurred due to noise interference in texture detection
Solution Approach 1:
The patent performs preliminary processing by generating multiple images at different resolutions and calculating texture determination values before final noise reduction. This preliminary action at reduced resolutions helps identify texture regions more reliably, preventing excessive blurring in subsequent processing steps.
Solution Approach 2:
The patent adds a new dimension of analysis by creating multiple images at different resolutions and computing texture determination values across these dimensions. This multi-resolution approach provides additional information for accurate texture detection, enabling better decisions about which regions require noise reduction versus which regions should preserve detail.
Data Source
AI summary
Provided herein is an electronic device and a method of image processing. An electronic device may include an image sensor configured to obtain image data including color pixel values and texture pixel values using color pixels and texture pixels, a resolution converter configured to generate two or more images having different sizes based on the image data, a parameter calculator configured to acquire two or more parametric images, respectively corresponding to the two or more images, using converted color pixel values and converted texture pixel values that are included in each of the two or more images, a parameter synthesizer configured to acquire a synthesized parametric image by synthesizing the two or more parametric images, and an image acquirer configured to acquire an output image based on the synthesized parametric image and the texture pixel values included in the image data.