Fast Image Segmentation Using Modified CLARA Algorithm

The CLARA algorithm is one of the popular clustering algorithms in use nowadays. This algorithm works on a randomly selected subset of the original data and produces near accurate results at a faster rate than other clustering algorithms. CLARA is basically used in data mining applications. We have used this algorithm for color image segmentation.The original CLARA is modified for producing better outputs. We used a technique of averaging of simulation results to reduce error due to sampling. We applied this algorithm for segmentation of color images due to the large amount of spatial coherency present in the image. The modified algorithm is also suitable for general data mining applications. From experimental results,we see that the suggested modification is a faster version of CLARA as well as able to produce better results.