Méthode d'évaluation univariante de la qualité des images. Application à des images comprimées selon la norme JPEG
Univariant assessment of the visual quality of images. Application to JPEG compressed images
OPEN ACCESS
This paper presents an original method to assess the quality of a grey level image. An example of application is presented on images compressed according to the JPEG standard. Contrary to most of the existing methods, this quality assessment is univariant, i.e. it doesn't require any reference image. The quality is represented by a mark whose variation depends on what use the image is for : visual, mathematical, data processing. A neural network is used to learn the murking way with a pool of known examples. The method i s then compared to classical bivariant methods to make sure that it i s reliable. The anticipated results arrived at have a less than 7 % precision.
Résumé
Cet article présente une méthode originale de détermination de la qualité d'une imageen niveaux de gris. Un exempled'application de la méthode sur des images comprimées selon la norme JPEG est présenté. Contrairement à la plupart des méthodes existantes, cette évaluation de qualité est univariante, c'est-à-dire ne nécessite aucune image de référence. La qualité est donnée sous la forme d'une note dont la progressionest étalonnée selon l'utilisation qui doit être faite de l'image : visuelle, mathématique, informatique. Pour ce faire, un apprentissage est effectué à l'aide d'un réseau de neurones sur une base d'exemples connus faite d'images dont on a préalablement noté la qualité avec le modèle souhaité. Pour s'assurer de sa fiabilité, la méthode est comparée à des méthodes bivariantes classiques. Elle permet de retrouver avec une erreur inférieure à 7 % les résultats prévus par celles-ci.
Image quality, ipeg, compression, block artefacts
Mots clés
Qualité image, ipeg, compression, artefacts de bloc
[ABP96] D. Anguita, I. Barbiéri. and F. Passagio. <<Block distortion assessment for image compression through ANNS>>. In Proceeding of the SPIE - The international society for optical engineering, volume 2664, pages 25-36, 1996.
[AMA95] S. Amari, <<The Handbook of Brain Theory and Neural Networks>>. MIT Press, Cambridge, Massachusetts, m. a. arbib edition, 1995.
[CAR93] R. Caruna. <<Generalization vs net size>>. Neural Information Processing Systems, 1993.
[CL93] H. Chou and Y.C.Li. <<A perceptually tuned subband image coder based on the measure of just-noticeable distortion profile>>. IEEE transactions on circuits and systems for video technology, 5(6): 467-476, 1993.
[DAL92] S. Dali. <<The visible differences predictor: an algorithm for the assessment of image fidelity>>. In Proceedings of the SPIE. Human Vision. Visual Processing and Digital Display III, volume 1666, pages 2-15, 1992.
[DAL93] S. Dali. <<The visible differences predictor: an algorithm for the assessment of image fidelity, in Digital Images and Human Vision>> MIT Press, a.b. watson edition, 1993.
[DC90] Recommandation 500-4 du CCIR. Méthode d'évaluation subjective de la qualité des images de télevision, 1990.
[EC96] T Eude and H. Chérifi. <<Optimisation perceptuelle d'un codeur jpeg>>. In Proc. of CORESA 96, pages 220-227, 1996.
[HUI75] C. Huijbregts. <<Regionalized variables and quantitative analysis of spatial data>>. In Display and analysis of spatial data. Wiley, j, davis and m. mac cullagh edition, 1975.
[KK93] S.A.Karunasekera and N.G.Kingsbury. <<A distortion measure for blocking artifacts in images based on human visual sensitivity>>. In Proceedings of the SPIE. Visual Communications and Image Processing, volume 2094, pages 474-486, 1993.
[KK94] S.A.Karunasekera and N.G.Kingsbury. <<A distortion measure for image artifacts based on human visual sensitivity>>. In Proceedings of the ICASSP-SPIE. volume 5, pages 117-120, 1994.
[KK95a] S.A.Karunasekera and N.G.Kingsbury. <<A distortion measure for blocking artifacts in images based on human visual sensitivity>>. IEEE Transaction on Image Processing, 4: 713-724, 1995.
[KK95b] S.A.Karunasekera and N.G.Kingsbury. <<A distortion measure for coding artifacts in images: Implementation aspects>>. Technical report, Department of Engineering, University of Cambridege, Signal Processing Group, 1995.
[LIN56] E.H.Linfoot. <<Transaction factors and optical design>>. Journal of the Optical Society of America, 46(9): 740-752, 1956.
[LIN58] E.H.Linfoot. <<Quality evaluation of optical systems>>. Optica Acta, 5(1-2): 1-13, 1958.
[MAA95] W. Maas. <<Vapnik-chervonenkis dimension of neural networks>>. In The Handbook of Brain Theory and Neural Networks, pages 522-52. MIT Press, Cambridge, Massachusetts, m. a. arbib edition, 1995.
[MAT71] G.Matheron, <<The theory of regionalized variables and its application>>. Les cahiers du centre de morphologie mathématique de Fontainebleau, 1971.
[MEC98] A.Mayache, T.Eude, and H.Chérifi. <<A comparison of image quality models and metrics based on human visual sensitivity>> . In Proceedings of the International conference on image processing, volume 3, pages 409-413, 1998.
[ML96] A.Meygret and D.Léger. <<In-flight refocusing of teh spot1 hrv cameras>> . In Proceedings of the SPIE, Algorithms for Multispectral and Hyperspectal Imagery II, vlume 2758, 1996.
[MS74] J.L.Mannos and D.J.Sakrison. <<The effects of visual fidelity criterion on the encoding of images>> . IEEE Transactions on Information theory, IT-20(4): 525-536, 1974.
[OLI98] J.L.Olives, <<Optimisation globale d'un système imageur à l'aide de critères de qualité visuelle>>. PhD thesis, ENSAE, 1998.
[RR89] M.Raffy and G.Ramstein. <<Analysis of the structure of radiometric remotelysensed images>>. Remote Sensing, 10(6): 1049-73, 1989.
[SHA56] O.H.Shade, <<Optical and photoelectric analog of the eye>>. Journal of the Optical Society of America, 46(9): 721-739, 1956.
[WAL91] G.K.Wallace, <<The jpeg still image compression standard>>. Communication of the ACM. 34:30-44,1991.
[WAT93] A.B.Watson, <<Visual optimization of dct quantification matrices for individual images>>. In Proceedings of AIAA Computing in Aerospace 9, pages 286-291, 1993.