Abstract
This paper presents a rule-based approach to classify natural and man-made fabric images. The interval data based on intensity values is used as feature vector. The mean and standard deviation of zero count and the maximum contribution in the intervals are computed. A rule base is developed, considering the mean and standard deviation of feature values. Eight varieties of natural and five varieties of man-made fabric images are considered. The classification is carried out at two levels. Firstly, the images are classified as natural fabric and man-made fabric images. At first level, the classification rates of 93% and 91% are obtained for natural and man-made fabric images respectively. In the second level, the individual fabrics are classified. The classification rates of 92.91% and 92.66% are achieved for varieties of natural and man-made fabric images respectively. The overall classification rate is found to be 92%.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have