Anurag Arnab

Papers

Total Citations

H-Index

About

Anurag Arnab is a leading researcher in computer vision, with a primary focus on video understanding, object segmentation, and efficient learning for real-world autonomous systems. His work bridges the gap between memory-augmented neural architectures and practical, few-shot visual recognition. In his highly cited 2023 paper, "Token Turing Machines," Arnab introduced a novel sequential Transformer model that incorporates an external memory of tokens to summarize historical context, enabling robust real-time visual understanding—a key contribution for applications like driverless cars and personal robotics. This work has already garnered 13 citations, reflecting its immediate impact on the field of sequential vision. Earlier, in his 2020 paper "Meta-Learning Deep Visual Words for Fast Video Object Segmentation" (9 citations), Arnab tackled the challenge of rapidly learning to segment novel objects in video without extensive fine-tuning, a critical capability for autonomous agents operating in unfamiliar environments. His research is distinguished by its elegant synthesis of meta-learning, memory mechanisms, and transformer architectures, consistently pushing the boundaries of how machines perceive and interact with dynamic visual scenes.