Image Processing Apparatus Periodic Noise Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods cannot accurately remove periodic noise components from images, as they often mistake subject frequency components for periodic noise, leading to incorrect removal of essential image features.
Innovation Solution
An image processing apparatus and method that reconstructs an image using a statistical model of the image structure, extracts periodic noise by calculating pixel differences, determines the noise frequency, and removes it, ensuring only unnecessary noise is eliminated.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If frequency processing is applied to remove periodic noise components, then periodic noise such as periodic unevenness and moiré can be removed, but subject frequency components may be mistakenly removed as well
Solution Approach 1:
The patent segments the frequency spectrum into multiple frequency bands and processes each band separately. By dividing the frequency processing into discrete bands, the system can identify and remove periodic noise in specific bands while preserving subject components in other bands, thus resolving the contradiction between noise removal and subject preservation.
Solution Approach 2:
The patent applies different processing characteristics to different frequency bands based on local properties. Each frequency band is analyzed and processed according to its specific characteristics, allowing targeted noise removal in bands containing periodic noise while maintaining subject integrity in bands containing only useful information.
2Object-affected harmful factors
If wavelet transform is used to reconstruct the image by nullifying signal components in frequency bands, then periodic noise can be removed, but judgment cannot be made as to whether noise components are periodic unevenness or moiré or subject frequency components
Solution Approach 1:
The patent employs dynamic analysis by examining the temporal and spectral characteristics of signal components across multiple frequency bands. Rather than static thresholding, the system dynamically adapts its processing based on the identified characteristics of each frequency band, enabling accurate differentiation between periodic noise and subject components.
Solution Approach 2:
The patent transitions from single-band frequency processing to multi-band frequency analysis, adding the dimension of frequency band segmentation. This dimensional expansion allows the system to identify and process periodic noise characteristics that are distributed across different frequency bands, improving the accuracy of noise identification and removal.
Data Source
AI summary
In order to accurately remove an unnecessary periodic noise component from an image, a reconstruction unit generates a reconstructed image without a periodic noise component by fitting to a face region detected in an image by a face detection unit a mathematical model generated according a method of AAM using a plurality of sample images representing human faces without a periodic noise component. The periodic noise component is extracted by a difference between the face region and the reconstructed image, and a frequency of the noise component is determined. The noise component of the determined frequency is then removed from the image.