Image Noise Reduction via Masked Pixel Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image compression techniques, such as JPEG, often degrade image quality by introducing noise around edges and text in digital images, especially when using lossy compression methods, and existing segmentation methods like MRC are complex and inefficient for reducing noise effectively.
Innovation Solution
A method that processes grayscale image data to identify noise-producing pixels and generates a mask to overlay and eliminate these pixels during output, allowing for efficient compression and storage while maintaining image quality, using a simpler segmentation technique that separates image data into planes for different compression methods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If lossy compression methods such as JPEG are used to reduce file size, then compression efficiency is improved, but image quality deteriorates due to noise around edges and text
Solution Approach 1:
The image data is divided into multiple planes or layers, with different compression methods applied to each plane. Critical regions (edges, text) are separated from non-critical regions, allowing selective compression that preserves quality where needed while achieving overall file size reduction.
Solution Approach 2:
Different compression qualities are applied to different regions of the image. High-quality compression is used for critical regions containing edges and text, while lower-quality compression is applied to regions where noise is less noticeable, optimizing the balance between file size and perceived image quality.
2Manufacturing precision
If complex segmentation methods like MRC are used to reduce noise, then image quality is improved, but processing complexity increases
Solution Approach 1:
The invention extracts only the essential information needed for noise reduction - specifically identifying edges and text regions - without implementing the full complexity of MRC segmentation. This selective extraction approach achieves noise reduction benefits while avoiding the computational burden of complete multi-plane segmentation.
Solution Approach 2:
The method performs preliminary identification of critical regions (edges, text) before compression, allowing subsequent noise reduction to be applied selectively. This preliminary action simplifies the overall process by avoiding the need for complex real-time segmentation during compression.
3Manufacturing precision
If reduction of compression amount is applied to reduce noise in JPEG compressed images, then image quality is improved, but file size increases
Solution Approach 1:
Instead of uniformly reducing compression across the entire image, the invention applies different compression levels locally - high compression to non-critical regions and low compression to critical regions containing edges and text. This maintains overall file size efficiency while improving quality where it matters most.
Data Source
AI summary
A system and method of reducing noise in output image data is provided. Grayscale image data having a plurality of pixels is received and processed. During processing, pixels which may produce noise are identified, and a mask associated with the image data is generated. The mask provides information related to the pixels, such as opaque and transparent regions for overlaying the pixels. The image data and the mask are compressed and stored. The mask assists in preventing the identified pixels from being visible when the image data is output, thereby reducing the noise in the image.