KUNWAR RANJEET
Biography
I am Dr. Kunwar Ranjeet, a Computer Science graduate from SRM Institute of Science and Technology, Kattankulathur, India. My work focuses on artificial intelligence, medical imaging, and explainable deep learning techniques. My recent research involves developing an interpretable skin cancer diagnosis framework using the Swin Transformer along with Grad-CAM and Grad-CAM++ methods, with the goal of enhancing early detection in dermatological oncology. I have also been a finalist in several national-level hackathons, and I am passionate about leveraging AI to build accessible and impactful healthcare technologies.
Research Interest
Abstract
XAI-Driven Skin Cancer Diagnosis
Skin is the largest organ in the human body and serves critical physiological and protective functions. Skin cancer, also referred to as cancer mortis, is among the most prevalent and rapidly increasing types of cancer worldwide. Timely and accurate diagnosis is essential for effective treatment. However, traditional diagnostic methods often rely heavily on expert interpretation, specialized equipment, and time-consuming procedures?factors that can delay early detection and treatment. To overcome these limitations, this study presents a deep learning-based skin lesion classification model utilizing the Swin Transformer architecture. This state-of-the-art model leverages a hierarchical structure and shifted window self-attention mechanism to capture both local and global features in dermatoscopic images. The model is trained on diverse datasets including HAM10000, ISIC-2008, ISIC-2019, and PH2, ensuring robust and generalizable learning. Crucial dermoscopic attributes such as pigment network and milia-like cysts were factored in during preprocessing and feature extraction, enhancing model interpretability and dermatological relevance. Explainable AI (XAI) techniques, including Grad-CAM and Grad-CAM++, are integrated to provide visual explanations that highlight important regions influencing the classification. The model's performance is evaluated using metrics such as accuracy, precision, recall, and F1-score. The results demonstrate strong classification capabilities and support the potential for early, AI-assisted skin cancer diagnosis. This work aims to bridge the gap between advanced AI and clinical usability, providing a transparent, accurate, and scalable solution for global dermatological applications.
