Image Processing Apparatus Motion Compensation Noise Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing techniques face challenges in accurately estimating motion vectors in noisy images, leading to block distortion and flickering/fluctuation phenomena in output images due to incorrect motion compensation and sudden changes between reference images before and after compensation.
Innovation Solution
An image processing apparatus with a motion estimation unit, motion compensation unit, flatness degree calculation unit, and pixel addition unit that calculates addition rates based on the flatness degree of the input image to accurately perform motion estimation and reduce block distortion by adding reference images before and after motion compensation, thereby minimizing sudden changes and improving image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If motion compensation is performed using conventional motion estimation in noisy images, then noise removal effectiveness is improved, but block distortion and flickering phenomena occur due to inaccurate motion vector detection
Solution Approach 1:
The patent performs preliminary noise removal processing on the input image and reference image before conducting motion estimation. By removing noise in advance, the motion estimation process operates on cleaner images, enabling accurate motion vector detection even in originally noisy conditions. This preliminary action resolves the contradiction by preparing the data in advance to ensure both noise removal effectiveness and motion detection accuracy.
Solution Approach 2:
The patent introduces an intermediate processing step where noise-removed images serve as mediators between the original noisy images and the motion estimation process. These intermediate images with reduced noise provide a cleaner basis for motion vector calculation, allowing accurate motion compensation without the block distortion and flickering that occur when estimating motion directly from noisy images.
2Reliability
If reference images before and after motion compensation are added to remove noise, then noise reduction is improved, but sudden changes cause flickering and fluctuation in output images
Solution Approach 1:
The patent dynamically adjusts the addition rate (blending ratio) of reference images before and after motion compensation based on the detected motion magnitude. For regions with large motion, it prioritizes reference images after compensation; for regions with small motion, it uses reference images before compensation. This dynamic adjustment prevents sudden changes and eliminates flickering while maintaining effective noise reduction.
Solution Approach 2:
The patent changes the parameter of addition rate (blending ratio) based on motion magnitude and flatness degree. By adjusting this parameter dynamically according to local image characteristics, the system smoothly transitions between using reference images before and after motion compensation, preventing sudden changes that would cause flickering while maintaining noise reduction effectiveness.
3Productivity
If motion estimation is performed on noisy images, then processing speed is maintained, but measurement precision deteriorates due to inability to discriminate signal from noise
Solution Approach 1:
The patent performs preliminary noise removal processing before motion estimation, preparing cleaner images in advance. This allows motion estimation to be performed accurately without requiring complex noise filtering during the estimation process itself, thereby maintaining processing speed while improving measurement precision.
Solution Approach 2:
The patent replaces the conventional approach of performing motion estimation directly on noisy images with a two-stage process: first noise removal, then motion estimation on cleaned images. This substitution of the processing mechanism eliminates the need for noisy-image motion estimation algorithms, achieving both speed and accuracy.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
To provide an image processing apparatus which reduces, using a temporal noise removal scheme, image block distortion or noise and temporal flickering and fluctuation phenomena. The image processing apparatus (1) removing noise in an input image by adding a reference image to the input image includes a motion estimation unit (20) which generates motion information indicating a motion estimated for the input image; a motion compensation unit (30) which generates a after motion compensation second image corresponding to a before motion compensation first image by performing motion compensation using the generated motion information; a flatness degree calculation unit (40) which calculates a flatness degree of the input image; a blending rate calculation unit (53) which calculates addition rates for the input image, first image, and second image, using the calculated flatness degree; and a pixel addition unit (60) which adds the respective images according to the addition rates.