ENHANCEMENT OF COLOR CONSTANCY USING TEXTURE PRESERVING SEGMENTATION METHOD

Author (s) : K.Gomathi and M.Sivabalan

---

Abstract :
The light sources influence the image and the uniform distribution of illuminant is necessary for an image. Color constancy is an important process in a number of vision tasks. It is violated by the presence of multiple light sources. Hence it is corrected using the color constancy algorithm which increases the performance by reducing the variations among the objects present in an image. In the existing method the algorithm is applied locally to the patches. It reduces the variations among similar and dissimilar objects. Experiments on spectral and real images showed that it reduces the influence of two light sources present in one scene. The segmentation based sampling used in the existing system slightly decreases the performance. In the proposed system texture preserving segmentation method is used. The qualitative analysis of the proposed system consistently enhances the image. This increases performance with distinct light sources. It preserves the boundaries and improves the visualization.

Keywords : Illuminant, color constancy, visualization, texture, sampling

Conference Date : 30/08/2019

Pages : 2

Paper ID : 19002