Noise Removal from Highly Corrupted Color Images with Adaptive Neighborhoods(Image)
スポンサーリンク
概要
- 論文の詳細を見る
A novel effective method for detection and removal impulse noise in highly corrupted color images is proposed. This detection-estimation method consists of two steps. Outliers are first detected using spatial relations between the color components. Then the detected noise pixels are replaced with the output of the vector median filter over a local spatially connected area excluding the outliers. Simulation results in a test color image show a superior performance of the proposed filtering algorithm comparing to the conventional vector median filter. The comparisons are made using a mean square error and a mean absolute error criteria.
- 社団法人電子情報通信学会の論文
- 2003-10-01
著者
-
Choi Tae-sun
Department Of Mechatronics Kwangju Institute Of Science And Technology
-
Kober Vitaly
Laboratory Of Digital Optics Institute For Information Transmission Problems
-
Choi Tae-sun
The Department Of Mechatronics Kwangju Institute Of Science And Technology
-
MOZEROV Mikhail
Laboratory of Digital Optics, Institute for information Transmission Problems
-
Mozerov M
Laboratory Of Digital Optics Institute For Information Transmission Problems
-
Mozerov Mikhail
Laboratory Of Digital Optics Institute For Information Transmission Problems
関連論文
- Fast Motion Estimation Techniques with Adaptive Variable Search Range (Special Section of Papers Selected from ITC-CSCC '98)
- Efficient File Management for Hard Disk Drive Embedded Digital Satellite Receiver(Regular Section)
- Fast Full Search Algorithm Using Adaptive Matching Scan Based on Gradient Magnitude
- Depth from Defocus Using Wavelet Transform
- Motion Estimation Based on Chain Code and Dynamic Programming(Fundamental Theories)
- Sampling Low Significance Bits Image to Reduce Quantized Bit Rate(Image Processing and Video Processing)
- Motion Estimation Based on Chain Code and Dynamic Programming
- Noise Removal from Highly Corrupted Color Images with Adaptive Neighborhoods(Image)
- Motion Estimation Based on Hidden Segmentation(Digital Signal Processing)