Student name:

Itai Levi

Project title:

Features extraction and classification of mammograms

Purpose:

One type of the abnormality in mammograms is a "mass". The purpose of this project was to extract features from ROIs (regions of interest) in mammograms, and apply classification methods on these features vectors in order to detect mammograms containing masses.

Algorithms:

  1. Feature extraction from Co-occurrence matrix in different directions and distances.

  2. Linear discriminant analysis of features vectors for classification.

References:

  1. Petrosian et al: Computer-aided diagnosis in mammography: classification of mass and normal tissue by texture analysis, Phys. Med. Biol. 39 (1994), pp. 2273--2288

  2. Sahiner et al: Computerized characterization of masses on mammograms: The rubber band straightening transform and texture analysis, Med.Phys.25 (4), April 1998, pp.516-526

  3. Heang-Ping Chan et al: Computer aided classification of mammographic masses and normal tissue: linear discriminant analysis in texture feature space, Phys.Med.Biol.40 (1995), pp.857-876