Advancements in Histopathologic Cancer Detection: A Deep Learning Odyssey
Kanwarpartap Singh Gill, Rahul Singh Chauhan, Nagendar Yamsani, Rupesh Gupta, Hani Mohammed Alshahrani, Adel A. Sulaiman, Mana Saleh Al Reshan, Asadullah Shaikh · 2024
This research paper focuses on the application of the proposed convolutional neural network (CNN) for histopathologic cancer detection, specifically targeting lymph node scans. Histopathologic cancer detection plays a crucial role in early diagnosis, and timely identification of metastatic tissues in lymph nodes is imperative for effective intervention and improved patient outcomes. The proposed methodology employs a deep learning model with a Sequential architecture, utilizing three sets of convolutional layers with increasing filter sizes, max-pooling layers, and dropout regularization to prevent overfitting. The model is practiced and checked on a smaller part of the PCam dataset. There are 220,025 images used for training and 57,468 images used for testing. These images were taken with a 10x zoom to see a larger area. The CNN model was tested and found to be really accurate at 92.72%. This shows that it can tell the difference between tumor and non-tumor classes really well. The research shows how important it is to use deep learning to help find cancer in tissue samples. It emphasizes how this can help diagnose and treat cancer early. The study also shows that the model used is good at accurately identifying different types of cancer. This research helps improve how we find cancer by using new technology that can make analysing tissue samples faster and better.