The aim of this paper is to introduce a robust CAD system that is able to increase the accuracy rate and reduce the false positive detection rate. This paper presents a system based on calculating the second order moment (variance) for the task of mass detection in digital mammogram. The goal is to develop a feature vector which is able to provide an accurate discrimination between the mass and normal tissues. The feature vectors are investigated in terms of their capability to achieve the classiﬁcation task using Random Forests with 10-fold cross validation. The proposed system has been tested using 1515 images from Image Retrieval in Medical Applications (IRMA) dataset and 265 images from Mammographic Image Analysis Society (MIAS) dataset. The study shows that the second order moment can be used efﬁciently for mammographic mass detection with accuracy of 100%.
Digital Object Identifier (DOI)
Meselhy Eltoukhy, Mohamed
"Mammographic Mass Detection Using Curvelet Moments,"
Applied Mathematics & Information Sciences: Vol. 11:
3, Article 10.
Available at: https://digitalcommons.aaru.edu.jo/amis/vol11/iss3/10