This document presents a novel hierarchical clustering algorithm for color image segmentation of Haematoxylin and Eosin (H&E) stained images of colon cancer. The algorithm aims to classify differences between benign and malignant tumor cells. It involves reading an input color image, coarse representation using 25 bins, hierarchical clustering to segment the image into classes, and outputting the segmented image. Intermediate steps include filtering, normalization, and segmentation of nuclei into a separate image in MATLAB for real-time simulation. Results show the algorithm improves color segmentation accuracy over previous k-means clustering approaches and provides features for region of interest classification of benign and malignant tissues.