Revolutionizing Cancer Diagnosis in IoMT with a Novel Lightweight Deep Learning Model for Histopathological Image Classification

Our proposed solution for breast cancer diagnosis offers a highly reliable and efficient alternative to the traditional manual examination method. By utilizing deep learning models designed with Convolutional Neural Networks (CNN), We have created a breast cancer detection system that is automated, inexpensive, and easy to transport. A system that significantly reduces the risk of delayed diagnosis and saves lives. This approach relies on an integrated ultrasensitive micro-bio-heat sensor array structured as a 3x3 grid using Altium software. The array functions as an embedded system aiming for the early identification of breast cancer. Leveraging IoT advancements, this system can connect with a server through smartphones. This study demonstrates the effectiveness of switch getting to know in histopathological photo classification. This research assesses the effectiveness of CNN models, both with and without transfer learning. It uses a pre-trained VGG16 model for image classification and demonstrates its successful implementation on a Raspberry Pi. This highlights its efficiency when running on a lightweight and portable processor. Our experimental results show that our system achieves an accuracy of 78% on the BreakHis database and can run on a Raspberry Pi device with minimal resources.