Title :
Mass detection in mammograms using ga based PCA and Haralick features selection
Author :
Amroabadi, SayedMasoud Hashemi ; Ahmadzadeh, Mohammad Reza ; Hekmatnia, Ali
Author_Institution :
Electrical and Computer Eng. Dept. University of Toronto, Toronto, Canada
Abstract :
Many existing researches utilized different types of feature extraction techniques to detect masses in ROI images. Based on our observations, inclusion of additional features beyond a certain point worsens the performance rather than enhancing it. This paper describes a hybrid method of mammogram recognition which is based on principle component analysis, Haralick features and Genetic algorithm to select the best features.
Keywords :
Algorithm design and analysis; Classification algorithms; Feature extraction; Genetic algorithms; Lesions; Principal component analysis; Support vector machine classification; Digital mammography; Genetic algorithm; co-occurrence matrices; component analysis;
Conference_Titel :
Electrical Engineering (ICEE), 2011 19th Iranian Conference on
Conference_Location :
Tehran, Iran
Print_ISBN :
978-1-4577-0730-8
Electronic_ISBN :
978-964-463-428-4
