TY - GEN
T1 - Restoration of original image from deteriorated image by probabilistic image model
AU - Karita, Yuji
AU - Tanaka, Toshiyuki
PY - 2008
Y1 - 2008
N2 - The conventional noise removal methods are based on spatial filtering and frequency filtering. But these methods have problems associated with degradation of image along side the noise removal. In this study, we propose the method that formulates noise based on multi-dimension Gaussian distribution and restore original image from deteriorated image by Probabilistic inference based on Bayesian statistics. The effectiveness of the proposed method has been validated using benchmark images.
AB - The conventional noise removal methods are based on spatial filtering and frequency filtering. But these methods have problems associated with degradation of image along side the noise removal. In this study, we propose the method that formulates noise based on multi-dimension Gaussian distribution and restore original image from deteriorated image by Probabilistic inference based on Bayesian statistics. The effectiveness of the proposed method has been validated using benchmark images.
KW - Bayesian statistics
KW - Gaussian white noise
KW - Image restoration
KW - Maximum likelihood estimation
UR - http://www.scopus.com/inward/record.url?scp=56749185384&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=56749185384&partnerID=8YFLogxK
U2 - 10.1109/SICE.2008.4655196
DO - 10.1109/SICE.2008.4655196
M3 - Conference contribution
AN - SCOPUS:56749185384
SN - 9784907764296
T3 - Proceedings of the SICE Annual Conference
SP - 3096
EP - 3100
BT - Proceedings of SICE Annual Conference 2008 - International Conference on Instrumentation, Control and Information Technology
T2 - SICE Annual Conference 2008 - International Conference on Instrumentation, Control and Information Technology
Y2 - 20 August 2008 through 22 August 2008
ER -