Breast Cancer Diagnosis from Histopathological Images of Benign and Malignant Tumors Using Deep Convolutional Neural Networks


Arslan A. B., ÇINARER G.

2nd International Congress of Electrical and Computer Engineering, ICECENG 2023, Bandirma, Turkey, 22 - 25 November 2023, pp.119-129 identifier

  • Publication Type: Conference Paper / Full Text
  • Doi Number: 10.1007/978-3-031-52760-9_9
  • City: Bandirma
  • Country: Turkey
  • Page Numbers: pp.119-129
  • Keywords: Artificial intelligence, Breast cancer, Deep learning
  • Yozgat Bozok University Affiliated: Yes

Abstract

Breast cancer is one of the popular types of cancer in women, so early diagnosis is particularly very important. Today, with the developing technology, many new methods are emerging for the accurate and rapid detection of this disease. Deep learning architectures developed with artificial intelligence technology have been widely used in this field in recent years. This study uses deep learning and image processing techniques to diagnose tumors from histopathological images using different and up-to-date deep learning architectures. Using the BreakHis dataset, the dataset is divided into a test training set and evaluated with five different deep learning algorithms. The performance of all algorithms in the study was evaluated and metrics such as F1 score and accuracy were obtained. The results show that the EfficientNetB0 model achieved the highest accuracy 91.59% and F1 score 90.40%. The results obtained show that deep learning algorithms can help in making accurate diagnoses of breast cancer and detecting the early stages of the disease. This study laid the foundation for deep learning algorithms to be tested on larger datasets and to be used in subsequent clinical studies.