Image segmentation.

The project is segmentation with superpixel clustering. By using SLIC Superpixels and boundary focused region margin. The program consists of three phase. First, generating superpixel by SLIC superpixel algorithm. It works by labeling every pixel in CIELAB and make a cluster. The second phase is to make a similarity matrix of these clusters. Each entry of matrix is the similarity between two superpixels. The similarity is the distance between the mean of every superpixel to its neighbors. Then combining the superpixels with the smallest distance together. Now we have regions. The final step is merging these regions by using threshold and the difference between adjacent regions in order to generate the segment and image segmentation will be done. The first step is done. From the second one, the part of calculating the mean of every superpixel is performed. The remaining parts, making a similarity matrix and find the smallest difference between superpixels and finally merging them together. I'm working with OpenCV. Thanks.