Bosom Malignant Diseases (Cancer) Identification by using Deep Learning Technique

Abstract

In recently observed that breast related diseases affects women present all over the globe, where it emerges as the second most common disease in the world. In 2012, 12 % cancer patients were present and from these patients 25 % are breast cancer patients. In the traditional method to cure the breast cancer is malignant tumor. Most of the doctors manually presumed the bosom malignant growth region. Various examinations have referred that this manual presumed requires more time and it relies upon the operation and machine. Therefore, it is necessary to design a perfect algorithm for the identification of bosom diseases. In this report, we have developed an algorithm to identify the breast cancer patient automatically. This algorithm can automatically detect the tumor of breast cancer by observing the biopsy pictures. Also, the calculation must be very precise, as the lives of individuals are at risk. All the performance operations are done on the microscopy pictures and the data set for this microscopy pictures is designed for the clustering analysis of a picture. The experimental results of the proposed scheme show accuracy 98.3 %, precision 0.65, Recall 0.95, F1 score 0.77 and ROC - AUC 0.692.