Involvement of Machine Learning for Breast Cancer Image Classification: A Survey

Table 16

SVM for breast image classification.

Reference	Descriptor	Image type	Number of images	Key findings
Taheri et al. [126]	Intensity information; Value of detected corner; Energy	Mammogram	600	Classified images into normal and abnormal images.; Removing unwanted objects from the images for reducing the redundancy and computational complexity.; Achieved Precision and Recall rates are 96.80% and 92.5%, respectively.
Tan et al. [127]	Shape, fat, presence of calcification texture, spiculation, Contrast, Isodensity type features selected; Total number of features 181	Mammogram	1200	Features have been selected from the region of interest.; They utilized the radial basis function (RBF) for their analysis.; The Sequential Forward Floating Selection (SFFS) method utilized for the feature selection.; The area under the receiver operating characteristic curve was (AUC) = .
Kavitha and Thyagharajan [128]	Histogram of the intensity has been used as a statistical feature.; 2D Gabor filter utilized for the textural feature extraction; Clinical features extracted from the database directly	Mammogram	322	When using SVM with the linear kernel the obtained Accuracy, Sensitivity, and Specificity are 98%, 100%, and 96%, respectively.; When using weighted feature SVM with weights the obtained Accuracy, Sensitivity, and Specificity are 90%, 100% and 75%, respectively.