Cite as:
Roubik K, Slaninka V, Dusek J, Klima M, Bernas M, 2002, "Model of the human visual system for the evaluation of perceptual image quality of compressed images" Perception 31 ECVP Abstract Supplement

Model of the human visual system for the evaluation of perceptual image quality of compressed images

K Roubik, V Slaninka, J Dusek, M Klima, M Bernas

Our aim was to create a model of the human visual system (HVS) suitable for evaluation of the perceptual quality of the compressed images, and to test the accuracy of the model by comparing the results predicted by it with those of the subjective image tests obtained from a group of observers. A special HVS model was developed, implemented, and tuned. The model comprises modelling of the optical part of HVS, colour separation with conversion into the CIE Lab colour space, Gaussian pyramidal decomposition, and brightness-contrast and colour-contrast processing simultaneously with masking and orientation-selective processing. The model works with 60 channels. Their outputs are weighted in order to compute a value representing the overall image quality. Evaluation of the accuracy of the model was carried out with subjective tests with a double stimulus continuous quality scale (DSCQS), conducted under standard conditions (ITU-R BT.500-6 recommendations). Perceptual image quality of images compressed by different methods and with different compression ratios was evaluated with the use of the model. The results are compared with the results of subjective tests. The excellent agreement obtained confirms suitability of the proposed model for objective image quality assessment of compressed images.

[Supported by grants GA CR No. 102/02/0133 and MSMT CR No. J04/98:212300014.]

These web-based abstracts are provided for ease of seaching and access, but certain aspects (such as as mathematics) may not appear in their optimum form. For the final published version of this abstract, please see
ECVP 2002 Abstract Supplement (complete) size: 1753 Kb