Shape-Based Tumor Retrieval in Mammograms Using Relevance-Feedback Techniques

Abstract. This paper presents an experimental "morphological analysis" retrieval system for mammograms, using Relevance-Feedback techniques. The features adopted are first-order statistics of the Normalized Radial Distance, extracted from the annotated mass boundary. The system is evaluated on an extensive dataset of 2274 masses of the DDSM database, which involves 7 distinct classes. The experiments verify that the involvement of the radiologist as part of the retrieval process improves the results, even for such a hard classification task, reaching the precision rate of almost 90%. Therefore, Relevance-Feedback can be employed as a very useful complementary tool to a Computer Aided Diagnosis system.

• Type: Journal
• Year: 2010
• Where: ICANN
• Authors: Stylianos D. Tzikopoulos, Harris V. Georgiou, Michael E. Mavroforakis, Sergios Theodoridis
• Keywords: Annotated Mass Boundary, Computer Aided Diagnosis, ICANN 2010, Neural Networks, Normalized Radial Distance