PreKiCan: A PARAM Utkarsh-Powered Model for Early Kidney Cancer Identification

The early detection of kidney cancer significantly increases the chances of better and successful treatment and obviously impacts patient’s survival. We conducted research that leverages the cutting-edge capabilities of the PARAM Utkarsh supercomputer, maintained by C-DAC Bangalore, to enhance the accuracy and speed of kidney cancer detection from x-ray images. We developed a deep learning binary classification model, that demonstrates a remarkable peak prediction accuracy of 99.8%, achieved through training with 250 epochs in high performance computing environment. The model incorporates the ultimate processing power of supercomputing to efficiently manage large datasets and perform complex computations, providing a remarkable improvement over traditional diagnostic method. This paper demonstrates the model development process within the high-performance computing environment and highlights the potential impact for the same in medical field, focusing the power of supercomputing technology in improving medical diagnostic processes and their outcomes in healthcare, to be more accurate.