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

Table 17

Bayesian classifier.

Reference	Descriptor	Image type	Number of images	Key findings
Kendall and Flynn [129]	Features extracted using DCT method.	Mammogram		Bayesian classifier obtained 100.00% sensitivity with 64.00% specificity.
Oleksyuk et al. [130]		—	—	Bayesian method obtained 86.00% with 80.00% specificity.
Burling-Claridge et al. [131]	Statistical and LBP features extracted.	Mammogram	322/410	Bayesian method obtained % and % Accuracy on MIAS and Inbreast image datasets (using statistical features).; Bayesian method obtained % and % Accuracy on MIAS and Inbreast image datasets (using LBP).
Raghavendra et al. [132]	Gabor wavelet transform utilized for feature extraction.	Mammogram	690	Locality Sensitive Discriminant Analysis (LSDA) for the data reduction.; NB obtained 84.34% Accuracy and 83.69% Sensitivity with 90.86% Specificity.
Pérez et al. [133]	23 features utilized.	Mammogram	—	UFilter feature selection methods utilized and its efficiency verified by Wilcoxon statistical test.
Rashmi et al. [134]	10 features utilized.	—	—	Benign and malignant tumors have been classified.
Gatuha and Jiang [135]	10 features utilized.	—	—	They built an android based benign and malignant tumor classifier.; Their obtained Accuracy is 96.4%