PEOPLE

Luca Lucchese, S.K.Mitra

OBJECTIVE

In this project we are investigating enhancement of color images. The objective of image enhancement is to process an image so that the result is more suitable than the original for a certain application. Usually, this enhancement consists in improving the image contrast, which is related to the sharpness of the details. In fact, textured details are more evident in images with high contrast, which, in turn, implies a large range in intensity and color variations. Contrast enhancement methods then aim at amplifying the local intensity or color variations within an image to increase the feature visibility. We are developing a technique for color contrast enhancement, which works in the xy chromaticity diagram associated with the CIE XYZ color space and consists of two steps. In the first step, all the "chromatic colors" are maximally saturated, i.e. moved to the borders of a selected color gamut. In the second step, the colors are desaturated according to the Center of Gravity Law for Color Mixture and by a suitable manipulation of the brightness level.

The idea behind this approach is simple: the colors of the surfaces of objects portrayed in a given image can be interpreted as stemming from the mixture of the "pure" colors intrinsic in the surfaces, and associated with their reflectances, and the color(s) of the illuminant(s). In the chromaticity diagrams, colors are added in compliance with the Center of Gravity Law, according to which the sum of two colors lies in a suitable location along the segment joining their chromaticities. If the desaturation operation is carried out towards the original white point, color enhancement is achieved. If instead a new white point with some chromatic bias is chosen, our technique is useful for color re-rendition because it is possible to simulate the change of the illuminant(s) of the scene. This functionality may be useful in applications concerned with image and video editing.

The ideas hinted by [1] have been further developed and extended to the CIE u'v' chromaticity diagram, which is perceptually more uniform than the xy. We are currently investigating the properties of this new class of color filters and their applicability to color image enhancement where they seem to offer great potentialities. 

A fourth line of our ongoing research is focused on color image equalization; we are developing a new technique for accomplishing this operation which is based on the representation of colors in terms of the CIE u'v' chromaticity coordinates and tri-stimulus value Y and on their separate equalization.

PUBLICATIONS

These materials are presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each authors copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.

1. L. Lucchese and S.K. Mitra, "Filtering color im­ages in the xyY color space," Proc. IEEE International Conference on Image Processing , Vancouver, B.C., Canada, September 2000, vol. III, pp. 500-533.