Automated Pathology Diagnosis Using Artifical Intelligence with Histopathological Images of Lungs

Lung cancer is of significant concern in India. Histopathological images are critical in the diagnosis of cancer and is usually done manually. Deep learning has a huge impact in the health care sector due to its high capabilities for accurate data analysis and early disease prediction. ResNet with predefined layers is used for diagnosis and 40 custom layers are added. The network is trained with images taken from the Kaggle data base. Both benign and malignant images are used for training. A Webpage is developed which serves as a graphical user interface for feeding the test images into the network. The network works on the test image and displays the results along with the type and stage. The network is robust with an accuracy of 98%.