---

Many early pattern recognition systems combined image measurements from different regions of the spectrum for classification. Until recently, however, color measurements were treated as generic features and the physical properties of color measurements were not exploited. With the onset of physics-based vision, researchers showed that color could be used to improve the capability of several processes in computer vision. These processes include segmentation, illumination-invariant recognition (color constancy), highlight interpretation, and interreflection analysis. Progress in these areas is summarized in references [1] and [2]. Current ICVL work in segmentation and color constancy is highlighted in this homepage. In the last few years, researchers have recognized the importance of image representations that characterize color texture. These representations are particularly important for modeling natural outdoor scenes. Early research in color texture analysis is summarized in reference [2] and current ICVL work is described in the texture analysis section of this homepage. Reference [3] presents physical and perceptual models for color and describes recent progress in color constancy and color segmentation.

[1] G. Healey, S. Shafer, L. Wolff (eds.)

Physics-based Vision: Principles and Practice, COLOR, Jones and Bartlett, Boston, 1992. introduction

[2] G. Healey

``Modeling Color Images for Machine Vision,'' in Advances in Image Processing and Machine Vision, J. Sanz editor, Springer-Verlag, 1996. abstract

[3] G. Healey and Q.-T. Luong

``Color in Computer Vision: recent progress,'' in Handbook of Pattern Recognition and Computer Vision, C.H. Chen, L.F. Pau, P.S.P. Wang editors, 2nd edition, World Scientific, 1997 abstract

[4] G. Healey

``Recognizing Objects in Color Images,'' in Proceedings of the Scandinavian Conference on Image Analysis, June 1997. abstract

Back to the ICVL homepage

Last modified: 10 May 1997