TY - GEN
T1 - Wavelets based on atomic function used in detection and classification of masses in mammography
AU - Juarez-Landin, Cristina
AU - Ponomaryov, Volodymyr
AU - Sanchez-Ramirez, Jose Luis
AU - Martinez-Reyes, Magally
AU - Kravchenko, Victor
PY - 2008
Y1 - 2008
N2 - Mammography is considered the most effective method for early detection of the breast cancer. However, it is difficult for radiologists to detect microcalcification (MC) clusters and camouflages masses. The mammograms (MG) images were decomposed into several subimages using Wavelet transform (WT) based on classical and novel class Wavelets using atomic function for reducing the volume of data in the classification stage. Various regions of interest (ROIs) in the MG images were selected where input data for multilayer artificial neural network (ANN) type classifier are formed applying the WT. We used different patterns to classify the normal, MC, spiculated and circumscribed masses ROIs. The detection performance has been evaluated on MG images from the Mammographic Image Analysis Society (MIAS) database. The proposed classification scheme was shown good performance in detecting the MC clusters and masses with acceptable classification.
AB - Mammography is considered the most effective method for early detection of the breast cancer. However, it is difficult for radiologists to detect microcalcification (MC) clusters and camouflages masses. The mammograms (MG) images were decomposed into several subimages using Wavelet transform (WT) based on classical and novel class Wavelets using atomic function for reducing the volume of data in the classification stage. Various regions of interest (ROIs) in the MG images were selected where input data for multilayer artificial neural network (ANN) type classifier are formed applying the WT. We used different patterns to classify the normal, MC, spiculated and circumscribed masses ROIs. The detection performance has been evaluated on MG images from the Mammographic Image Analysis Society (MIAS) database. The proposed classification scheme was shown good performance in detecting the MC clusters and masses with acceptable classification.
KW - Atomic function
KW - Classification
KW - Mammography
KW - Microcalcification
KW - Neural networks
KW - Wavelet transform
UR - http://www.scopus.com/inward/record.url?scp=57049181231&partnerID=8YFLogxK
U2 - 10.1007/978-3-540-88636-528
DO - 10.1007/978-3-540-88636-528
M3 - Contribución a la conferencia
SN - 3540886354
SN - 9783540886358
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 295
EP - 304
BT - MICAI 2008
T2 - 7th Mexican International Conference on Artificial Intelligence, MICAI 2008
Y2 - 27 October 2008 through 31 October 2008
ER -