An Approach to Color Image Quality Objective Assessment by Combining Human Vision Based General Perceptive Deviation and r-g Chromatic Peak Value Offset

To meet the requirement of color image quality assessment in applications of the image color restoration, the index of histogram perceptive deviation evaluation, based on the histogram cumulative distributions of images' grey information, is generated by introducing such correction factors as human vision sensitivity, etc. Then general perceptive deviation algorithm is proposed via combining histogram perceptive deviation with entropy and average gradient values. A novel no-reference objective image quality assessment approach is achieved by associating general perceptive deviation parameter with r-g chromatic peak value offset parameter. To test the performance of the algorithm above, a standard color image database under different illumination conditions is established. Compared with the usual objective evaluation indexes, the proposed approach is more coincident with the subjective assessment method in the brightness perception, hue value and synthetic quality.