Most objective image evaluation procedures employ distortion and noise measures that are mathematically simple, such as mean square error. However, these measures do not usually give the same results as subjective evaluation of image quality. Results of experiments measuring the human visual system response to noise patterns containing various spatial frequency contents and structures are described in this paper. These results can be used to postulate mathematical models for the subjective effect of pictorial noise. The hope is that noise effects calculated using these models will correlate reasonably well with actual subjective effects, thus allowing design and optimization procedures to use a quantitative measure of subjective effects.
|Original language||English (US)|
|Number of pages||8|
|Journal||Photogr Sci Eng|
|State||Published - Jan 1 1977|
ASJC Scopus subject areas