Project: Color Image Segmentation

PEOPLE

L.Lucchese, S.K.Mitra

OBJECTIVE

In this project we have investigated both supervised and unsupervised color image segmentation problems, and have developed a number of new algorithms.

ALGORITHMS

One of the algorithms developed for supervised segmentation of color images is based on the palletized representation of the image and consists of the following steps: down-sampling the lattice upon which the image pixels are de­fined, blurring the image through low-pass filtering in order to average colors lying in almost uni­form regions, exploiting the well-established and fast methods for color quantization in order to compute color clusters in the RGB space.  Computer simulations have verified the perfor­mance of the new method.  A paper describing this method was presented at an international conference [6].

An algorithm developed for unsupervised color image segmentation is based upon the following ideas: processing images in their palletized format; using the low spatial frequency content in the low-low band of the 2-level wavelet transform of the image; building and thresholding hue and chroma histograms in the CIELUV uniform color space; rec­ognizing the main colors within an image by successively estimating their hue, chroma and lightness coordinates; and matching color palettes in the CIELUV space.  Extensive computer simulations have verified the effectiveness of this new approach.  Results of this investigation have been presented at a conference [5].

Two feature-based segmen­tation techniques were developed that, wherever possible, handle color images in a palletized format to speed up the computation time.  The first algorithm achieves seg­mentation by spanning and clustering colors in the CIELUV space, first hue-wise and then within planes containing chroma and lightness coordinates; one relies on the low-pass content of the image's wavelet decomposition and uses his­tograms.  A paper describing this method was pre­sented at a conference [4].  The second algorithm follows instead a different strategy: colors are first clustered in a 2D chromatic space and the third coordinate represented by lightness is found in a second stage.  Results of this project were presented at an international conference [3]. Another algorithm for color image segmentation has been developed; it is based on separately filtering with anisotropic diffusion the chromatic and achromatic channels of the image, which are then separately segmented.  These new results will be presented at an international conference on image processing [2].  An expanded version of this work has recently been accepted for publication in a journal [1].

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, " Color segmentation based on separate anisotropic diffusion of chromatic and achromatic channels,"  IEEE Proceedings: Vision, Image, and Signal Processing, 2001 - to be published.

2. L. Lucchese and S.K. Mitra, "Color image segmentation through independent anisotropic diffusion of complex chromaticity and lightness,"  Proc. IEEE International Conference on Image Processing (ICIP 2001), Thessaloniki, Greece, September 2001 - to be published.

3. L. Lucchese and S.K. Mitra, "Unsupervised segmentation of color images based on k-means clustering in the chromaticity plane," Proc. of IEEE Workshop on Content-based Access of Image and Video Libraries (CBAIVL'99), Fort Collins, CO, June 1999, pp. 74-78.

4. L. Lucchese and S.K. Mitra, "Unsupervised low frequency driven segmentation of color images," Proc. 1999 IEEE International Conference on Image Processing , Kobe, Japan, October 1999, vol. III, pp. 240-244.

5. L. Lucchese and S. K. Mitra, "Unsupervised color image segmentation," Proc. IEEE  Workshop on Multimedia Signal Processing , Los Angeles, CA, December 1998, pp. 33-38.

6. L. Lucchese and S. K. Mitra, "An algorithm for fast segmentation of color images, " Proc. IEEE 10-th Tyrrhenian Workshop on Digital Communication , Ischia, Italy, September 1998, pp. 110-119.

Back to Image Segmentation