Enhanced Detection and Retrieval of Ocular Melanoma Using Pattern Recognition

Ocular Melanoma is the most common primary eye tumor among adults in United States with around 2500 new cases reported each year. With the increasing number of cases and the aggressive growth of disease, the need arises to cut down the time spent on analyzing hundreds to thousands of ultrasound scans, MRI and CT scans images each day. The proposed enhanced detection and retrieval of ocular melanoma using pattern recognition involves three main steps: feature detection and feature extraction and similarity measures. The proposed technique uses integrated features for feature detection and retrieval of ocular melanoma. The texture features are extracted using Gray level co-occurrence matrix (GLCM) technique. And the color features are extracted using dispersion methods. The distance between the query image features and the features of database images is computed using Euclidean distance. Precision and Recall are used for analysing the performance of the algorithm. Precision is the measure of ability of a system to present all the relevant items and Recall is defined as the ability to present only relevant items.