Analysis of breast lesions using laws' mask texture features
Sahil Bhusri, Shruti Jain · 2016
Breast Cancer is one of the most lives threatening cancer in women, the earlier diagnosis of cancer is always a big plus in the treatment. The region suffering from damage is known as lesion and the breast lesions are classified into two categories i.e. benign and malignant. In this work the analysis of breast lesions is completed using the texture features. Laws' mask texture features of different dimensions i.e. of dimension 3, 5, 7 and 9 are computed from the ultrasound images of the breast lesions. These Laws' masks are used to compute various statistical features (Energy, Kurtosis, Mean, Standard Deviation and Entropy). SVM classifier is used for classification and an overall classification accuracy of 97.4% and the individual classification accuracy values of 90.4% and 100% have been yielded for benign and malignant classes respectively.