Title :
Cascaded Neural Network-based S-VHS Restoration
Author :
Le Dinh, Kha ; Le Dinh, Chon-Tam
Author_Institution :
Univ. of Sherbrooke, Sherbrooke
fDate :
5/1/2007 12:00:00 AM
Abstract :
A restoration method for S-VHS sequences is proposed in this paper to help to improve their image quality. The solution consists of a cascaded neural network structure using restored luminance signal. The proposed method was tested with various color images and experimental results showed that both subjective and objective image quality were substantially improved.
Keywords :
home automation; image colour analysis; image restoration; neural nets; video signal processing; S-VHS restoration; cascaded neural network; color images; image quality; image restoration; objective image quality; restored luminance signal; super video home system; Deconvolution; Frequency modulation; Image quality; Image restoration; Magnetic recording; Neural networks; Optical recording; Radio spectrum management; Signal restoration; Video recording;
Journal_Title :
Consumer Electronics, IEEE Transactions on
DOI :
10.1109/TCE.2007.381723
