GPU Based Digital Histopathology and Diagnostic Support System for Breast Cancer Detection: A Comparison of CNN Models and Machine Learning Models
Mradul Kumar Jain1, Nirvikar2, Amit Kumar Agarwal3
1Mradul Kumar Jain, Department of Computer Science & Engineering, ABES Engineering College, Ghaziabad (U.P), India.
2Dr. Nirvikar, College of Engineering Roorkee, Roorkee (Uttarakhand), India.
3Amit Kumar Agrawal, ABES Engineering College, Ghaziabad (U.P), India.
Manuscript received on 18 April 2019 | Revised Manuscript received on 25 April 2019 | Manuscript published on 30 April 2019 | PP: 367-376 | Volume-8 Issue-4, April 2019 | Retrieval Number: D6151048419/19©BEIESP
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: One of the decisive reasons of cancer is uncontrolled augmentation of cancerous cells, malignant cells, or tumor cells in any living organism life at any stage. The scientific role of pathology is to diagnosis and prognosis of diseases to find changes at the level of cell structures with cell components including nucleolus and cytoplasm, tissues (i.e. grouped cell with complex structures) and organs which in turn give rise to the presenting signs and symptoms of the patient. It has been observed by clinical pathology system and histopathology methods that damaged or unrepaired cells do not die and show uncontrolled growth – a reason to mass development of cancerous cells. Frequently, cancerous cell travel through the blood and lymph systems, and cross the effected boundary organs to other body region where they repeats the process of uncontrolled growth cycle. This process of cancer cells leaving one region and growing in other part of body system is termed as metastatic spread or metastasis. Histopathological methodology can detect breast cancer. This diagnosis can be done with various Machine Learning Models and Deep Learning based Convolutional Neural Networks Models. The analysis shows convolutional neural networks models provide significant accurate results in comparison machine learning based models.
Keywords: Digital Histopathology Images Processing System (DHIPS), Nonlinear Mapping, Principal Component Analysis, CNN (Convolutional Neural Networks),Visual Geometry Group (VGG), Alexnet, Googlenet
Scope of the Article: Images Processing