PROBABILISTIC MEASURE OF COLOUR IMAGE PROCESSING FIDELITY

Eugeniusz Kornatowski - Krzysztof Okarma

   In the paper a probabilistic approach to quality assessment of image processing algorithms is proposed. Presented scalar measure can be used for any colour space and gives very similar results regardless on the image content. It can be an interesting supplement to existing image quality metrics in applications where the details of the processing algorithm are known. Its good correlation with subjective evaluation is comparable to some other recently proposed measures such as Universal Image Quality Index, Structural Similarity and SVD-based measure.

Keywords: colour image quality assessment, image processing fidelity, statistical image processing, objective measures

[full-paper]