The property of fabrics to release stains is an important quality that impacts the pricing of the fabric on the marketplace. Samples of fabrics stained with common household products, such as ketchup, mustard, ragu and corn oil, are prepared and the degree of stain release is evaluated visually with respect to a standard set of AATCC stain release replicas. However, the current method of manual grading for stains is subjective, cumbersome, and leads to inaccuracies due to human inconsistency. The current work is an effort to develop a novel and automated system based on an image analysis approach for objective evaluation of stain release. The fabrics are converted to digitized images using a flat-bed scanner and passed to the input of the image analysis system. This system has the ability to automatically detect the presence of stains on a fabric image and, if a stain is detected, the boundary of the stain is automatically extracted and a suitable grade (that is in keeping with the AATCC grade) is assigned to the stain. In addition to the existing grade, additional color measurements that characterize the stain are also extracted from the fabric images. The performance of the system was validated on fabric datasets prepared at the International Textile Center. The overall success rate of the system in detecting stains is currently at 90%, and the preliminary results look promising.