Screening is essential for extraction of object Candidate in Diabetic Retinopathy (DR).Purpose is to detect aberrant from color fundus images. Pixels with number of lesions can be illuminated by region Segmentation Using color space selection. In the proposed method simple color detection in LAB color space has done. Four color models, like RGB, Luv, Lab and HVC, are calculated which uses the object-based color difference image. The objectbased color difference images of bright and dark objects are detected respectively. This helps in objects detection in Diabetic retinopathy like exudates hemorrhages and blood vessel. Thus by computing the color difference we can easily classify objects. Such a screening systems may reduce efforts of ophthalmologist. In this paper our goal is to analyze and evaluate the various color spaces in color image enhancement applications. Conversion accuracy and similarity measure are the two objective parameters to measure the performance of each color space [1].