Jogendra Nath Kundu

Papers

Total Citations

H-Index

About

Jogendra Nath Kundu is a researcher whose work sits at the exciting intersection of computer vision, graphics, and human-robot interaction. His key research areas include sketch recognition, human motion modeling, and generative sequence modeling. Kundu’s most notable contribution is his pioneering work on enabling robots to understand freehand sketches, as detailed in his highly cited 2016 paper "Enabling My Robot To Play Pictionary" (47 citations). In this work, he proposed a recurrent neural network (RNN) architecture that exploits the sequential, stroke-by-stroke nature of sketching—a significant advance over prior methods that treated sketches as static images. This approach not only improved object recognition accuracy but also laid the groundwork for more intuitive human-robot communication. Building on this, Kundu tackled the complex challenge of modeling inter-person human motion interactions, introducing cross-conditioned recurrent networks for long-term synthesis of realistic, multi-person motion sequences (21 citations). His work has implications for animation, surveillance, and robotics, demonstrating a consistent focus on capturing temporal dependencies in sequential data. Through these contributions, Kundu has established himself as a leading voice in sequence-based visual understanding.