Papers
22
Total Citations
652
H-Index
11
About
Sourav Garg is a prominent robotics and computer vision researcher whose work sits at the intersection of visual place recognition (VPR), semantic mapping, and robot navigation. He is perhaps best known for his contributions to making robots reliably recognize and understand their environments under challenging, real-world conditions — across varying lighting, weather, and viewpoints. His landmark paper *AnyLoc* (2023, 163 citations) pushed toward universal VPR systems capable of generalizing beyond specific environments or tasks, a longstanding bottleneck in the field. Complementing this, his widely-cited survey on semantics for robotic mapping, perception, and interaction (2020, 100+ citations) provided the research community with a foundational reference for understanding how semantic understanding enables richer robot-world interaction. His work on semantic-geometric VPR (2019, 98 citations) introduced fresh perspectives on reconciling competing approaches to place recognition. Beyond navigation, Garg has contributed tools like *OpenSeqSLAM2.0* for benchmarking VPR systems, explored robot-based retail stock assessment, and more recently advanced language-grounded robot planning through *SayPlan* and topological mapping via *RoboHop*. His research consistently bridges theoretical advances with practical robotics applications, making him a valuable and versatile voice in the autonomous systems community.
Research Focus
Key Achievements
Top Papers
- 1<i>AnyLoc</i>: Towards Universal Visual Place Recognition163 citations · 2023
- 2Semantics for Robotic Mapping, Perception and Interaction: A Survey100 citations · 2020
- 3
- 4Visual Place Recognition: A Tutorial [Tutorial]53 citations · 2023
- 5Semantics for Robotic Mapping, Perception and Interaction: A Survey44 citations · 2020
- 6
- 7
- 8
- 9High-fidelity simulation for evaluating robotic vision performance20 citations · 2016
- 10