Novel Parameter-less Hierarchical Algorithms for Image Segmentation and Data Clustering

This thesis presents novel algorithms for image segmentation and data clustering. A segmentation algorithm partitions an image into non-overlapping blocks and a clustering algorithm finds the natural grouping in a dataset. The proposed algorithms are hierarchical and require very few if any parameters. As hierarchical algorithms, these methods provide multiple solutions to segmentation or clustering and an appropriate solution can be chosen depending on the application requirements. Our algorithms are memory and time efficient and outperform majority of the existing methods. This thesis also demonstrates an application of our segmentation algorithm to detect noise in images.