Advances in Machine Learning Techniques for Sign Language Interpretation
Published in Springer LNNS (ICTCS 2024) - survey of CNN, LSTM, Transformer, and hybrid methods for sign language interpretation.
Explore featured publications, research, and projects below, or see them all here.
Published in Springer LNNS (ICTCS 2024) - survey of CNN, LSTM, Transformer, and hybrid methods for sign language interpretation.
Final Year Project - real-time ISL-to-text using MediaPipe landmarks, achieving 94.32% accuracy on dynamic gestures with DTW.
A software raytracing engine written entirely in Java with no external libraries. Supports primitives, OBJ models, coloured lighting, texturing, normal mapping, and reflection.
Independent Study (Honours Seminar) - lightweight CNN for rain-streak removal, outperforming baselines on PSNR and SSIM.
Real-time 2D Navier-Stokes fluid simulation in pure Java with interactive dye injection and velocity visualization.
Voxel-style assets in Blender, wave-based enemy system, placeable barrier mechanics, and Oculus performance optimizations. Built during a 50-hour AR-VR workshop at PICT in collaboration with Appy Monkeys Pvt. Ltd.
Hewlett Packard Enterprise (via Mphasis Ltd.) • Pune, India
Hewlett Packard Enterprise (via Mphasis Ltd.) • Pune, India
MITU Skillologies • Pune, India
PICT • 2022 – 2025