Abstract

ADVANCED DETECTOR FOR INVASIVE DUCTAL CARCINOMA USING DEEP LEARNING

N. L. Anbarivan*, V. Geethanjali and V. Yeshwanth

Abstract

The utilization of deep learning intelligence system to redefine medical assistance and support for decision making is shifting the paradigm of disease diagnosis. Clinical grade systems that are able to automate the process of detection of disease will help in effective decision making. Traditional methods of clinical screening and analysis suffers from various complexities varying from specificity to image noise. The objective of this research is to utilize deep learning intelligence system to redefine medical assistance and support for decision making and shift the paradigm of disease diagnosis. Clinical grade systems that are able to automate the process of detection of disease will help in effective decision making. We propose a reliable model that is able to capture cancer cells in images and support in making the process of treatment faster. Neural networks and transfer learning has been employed to detect the presence of cancer cells. In addition, automatic feature extraction and classification has been exploited to support the objective. We have achieved an accuracy of 95% in the test set and 94% in the training set and we have taken a step ahead to transform the idea of accessing care using our unique medbots.

Keywords: Breast cancer, intelligence prognostic tool, retrained models, InceptionV3, Automatic feature extraction.