Custom Convolution Neural Network for Breast Cancer Detection
Thyagaraj T1, Keshava Prasanna2, Hariprasad S A3
1Thyagaraj T, Department of Electronics and Communication, BMS Institute of Technology and Management, Visvesvaraya Technological University, Belagavi, India.
2Keshava Prasanna, Department of Horticulture, Keladi Shivappa Nayaka University of Agricultural and Horticultural Sciences, Shivamogga (Karnataka), India.
3Hariprasad S A, Faculty of Engineering and Technology, Jain Deemed to be University, Bengaluru (Karnataka), India.
Manuscript received on 24 November 2023 | Revised Manuscript received on 01 December 2023 | Manuscript Accepted on 15 December 2023 | Manuscript published on 30 December 2023 | PP: 22-29 | Volume-13 Issue-2, December 2023 | Retrieval Number: 100.1/ijeat.B43341213223 | DOI: 10.35940/ijeat.B4334.1213223
Open Access | Editorial and Publishing Policies | Cite | Zenodo | 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: Breast cancer remains a serious global health issue. Leveraging deep learning techniques, this study presents a custom Convolutional Neural Network (CNN) framework for breast cancer detection. With the specific objective of accurately classifying breast cancer, a framework is developed to analyse high-dimensional medical image information. CNN’s architecture, which consists of specifically developed layers and activation components tailored for breast cancer categorisation, is described in detail. Utilising the BreakHis dataset, which comprises biopsy slide images of patients at various cancer stages, the model is trained and validated. Comparing our findings to those of conventional techniques, we observe notable improvements in sensitivity, specificity, and accuracy. Gray-Level Co-Occurrence Matrix (GLCM) features extracted from the BreakHis dataset were used to analyse the performance of sequential neural network, transfer learning, and machine learning models. After analysis, we have proposed hybrid models of CNN-SVM, CNN-KNN, CNN-Logistic regression and achieved an accuracy of about 95.2%
Keywords: Breast Cancer Detection, CNN, Mobile Net
Scope of the Article: Convolutional Neural Network