| 摘要 |
PROBLEM TO BE SOLVED: To learn a coefficient and to generate optimal data using the coefficient, in order to further enhance a resolution, S/N, compression distortion, luminance value and the like of an input image signal. SOLUTION: A processing circuit 3 applies image quality reduction processing such as processing f(x) like LPF, compression/decompression, gain and offset, for example, to an input image signal. A class code (index) is generated from the image signal to which processing f(x) has been applied, and the generated class code is supplied to a learning circuit 7. Switches 4 and 6 are changed over to supply, to the learning circuit 7, the image signal to which processing f(x) has been applied and a teacher signal, the image signal to which processing f(x) has been applied, and the input image signal, and the input image signal and the teacher signal. In the learning circuit 7, in order to acquire coefficients of a linear combination expression from these signals, learning is performed in accordance with least square error standards. COPYRIGHT: (C)2008,JPO&INPIT |
