Harikrishnan Raghukumar

   Through this project I tried to create a prototype of how Cancer diagnosis can be implemented through Telepathology - which will reduce the workload of pathologists using advancements in Deep Learning.

   For this, I primarily built a U-Net based deep neural network which can detect and segment lesions from Whole Slide Images (WSI). (lesion is an abnormal area of tissue, which has an increased risk of developing into cancer. Early detection of such cells leads to significant improvement in patient survival rate.) On top of it, I implemented active learning and deployed the network in a server using flask to make it accessible from anywhere across the world.

   Further details are present in the below slide