Open AccessTexture Analysis of Mammogram Using Histogram of Oriented Gradients Method
Author(s) -
Athraa H. Farhan,
Mohammed Y. Kamil
Publication year - 2020
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/881/1/012149
Subject(s) - histogram , adaptive histogram equalization , histogram equalization , breast cancer , pattern recognition (psychology) , artificial intelligence , mammography , contrast (vision) , computer science , feature (linguistics) , computer aided diagnosis , support vector machine , texture (cosmology) , computer vision , medicine , cancer , image (mathematics) , linguistics , philosophy
The second foremost reason for dying ladies all across the world is breast cancer. The possibilities of survival can be raises when cancer detects earlier; therefore, the mortality reduction. The radiologist used mammograms to recognize breast tumors at an early level. Since the mammograms have little contrast, hence, it is unclear to the radiologist to distinguish small tumors. A computer-aided detection system contributes to explaining mammograms and helps the radiologist to indicate the appearance of breast mass and discriminate among normal and abnormal tissue. In this research, we introduce a histogram of oriented gradients as a method of feature extraction and identify mass regions in mammograms. The features extraction from this method classified by a support vector machine. To improve the diagnosis ability, contrast limited adaptive histogram equalization pre-processing system is utilized. Mini-MIAS database used to estimate this approach. The top accuracy, sensitivity, and specificity obtained are 90%, 69%, 100%, respectively.