Extraction of Orientation Field Using Gabor Filter and Gradient Based Approach for the Detection of Subtle Signs in Mammograms

摘要

Architecture Distortion is a very significant finding in determining breast abnormality found among women's. But due to its varying nature, it is considered very difficult to detect as compared to other abnormalities like masses and micro-calcifications. The purpose of this study is to develop a method for detection of architecture distortion by using a technique comprising of Gabor filter and Gradient based method. The proposed method is used to characterize the oriented patterns, which are indicative of distortion in normal breast patterns. The image database consists of 60 images from two standard data sets (Mammogram Image Analysis Society (MIAS): 19 cases with architecture distortion and 16 normal cases, Digital Database for Screening Mammography (DDSM):15 cases with architecture distortion and 10 normal cases). For subjective analysis the orientation field is analyzed for a mammogram showing both the normal and abnormal areas. While for objective analysis, entropy features are deduced from both the normal and abnormal mammogram. An Accuracy of 100% is found with objective analysis and 93.33% through subjective analysis, with combined data set. It was concluded that proposed technique were effective to detect architectural distortion.