Applying the Deep Learning Model on an IoT Board for Breast Cancer Detection based on Histopathological Images
Shahirah Binti Zahir, Amiza Amir, Nik Adilah Hanin Zahri, Wei Chern Ang
Abstract
Abstract In breast cancer diagnosis, pathologists evaluate microscopic images of tissue samples to determine if it is benign or malignant. The manual examination process could result in delayed diagnosis, which leads to late cancer treatment and can risk lives. In this paper, we proposed an automated, low-cost, and portable breast cancer detection based on histopathological images by using deep learning. The deep learning models were designed by using the Convolutional Neural Network (CNN). This paper compares the performance of the CNN model by using transfer learning utilizing a pre-trained model (VGG16) and the performance of a CNN model without transfer learning. The result shows that transfer learning provides a good base for classification of histopathological images. The model was successfully deployed on a Raspberry Pi, which demonstrates the model efficiency to run on a lightweight and portable processor.