作者: Rafael Llobet , Roberto Paredes , Juan C Pérez-Cortés , None
DOI: 10.1007/11492542_61
关键词:
摘要: Although screening mammography is widely used for the detection of breast tumors, it difficult a radiologist to interpret correctly mammogram. It possible improve this task by using computer aided diagnosis system (CAD) which highlights areas most likely contain cancer cells. In paper, we present and compare five different feature extraction methods in digitized mammograms. All are based on features extracted from local window k-nearest neighbor classifier with fast search.