Graphics and Media Lab

Object brightness estimation is an important task at the stage of quality control in manufacturing of, e.g., lighting products, paintwork materials, interior items, etc. Thus in addition to the direct evaluation of reflective and light-emitting material properties visual evaluation happens to be essential. So the properties of the material should be assessed in terms of human brightness perception, that is, we need estimation of object?s perceived brightness. In this paper, we propose a solution to the problem of object?s integral perceived brightness estimation using high dynamic range images as input. High dynamic range image contains luminance values of physical objects, so considering them allows obtaining a more correct result, compared to working with single low dynamic range image. In general, the proposed method consists of two steps: the global and the local transformation of brightness. Both steps are based on psychophysical researches and increase the precision. Experimental verification has shown the superiority of the proposed method of integral subjective brightness estimation over several simpler methods, including simple averaging of luminance.