Image Processing Apparatus Noise Motion Differentiation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional noise determination methods, such as Mean Absolute Difference (MAD), cannot distinguish between noise interference and object movement in video content, leading to incorrect noise suppression and potential image streaking.
Innovation Solution
An image processing apparatus and method that calculates pixel differences and edge types between frames, determines a noise level by comparing edge types and pixel differences, and adjusts image blending based on the noise level to differentiate between noise and motion, thereby generating an output image that accurately represents the noise level.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Mean Absolute Difference (MAD) is used to determine noise level, then noise detection capability is improved, but inability to distinguish motion from noise causes image streaking and reduces measurement precision
Solution Approach 1:
The patent segments the image into multiple blocks and further divides each block into sub-blocks for edge detection. By processing different regions separately and analyzing edge characteristics in each sub-block, the method can distinguish between noise-induced pixel changes and motion-induced pixel changes, preventing image streaking while maintaining noise detection accuracy
Solution Approach 2:
The patent introduces edge detection as an intermediary mechanism to mediate between pixel difference calculation and noise level determination. By using edge information as an intermediate indicator, the system can differentiate whether pixel value changes are caused by noise or motion, thereby improving both measurement precision and reliability
Data Source
AI summary
An image processing apparatus for processing a previous image having first pixels and a present image having second pixels. The image processing apparatus includes: a pixel difference calculating unit which calculates pixel differences between corresponding pairs of the first and second pixels, and outputs pixel difference values; an edge processing unit which detects and compares edge types of the first and second pixels, sums a number of the edge types that are the same, and outputs a sum value; a noise level processing unit which calculates a noise level of the present image according to the sum value and the pixel differences; a blending value determining unit which determines a blending value according to the noise level; and an output unit which adds weights of the present and previous images according to the blending value and outputs an output image. An image processing method is also disclosed.