Aravindhan K Krishnan

Indian Institute of Technology Hyderabad, Amazon (United States)

Papers

Total Citations

H-Index

About

Aravindhan K Krishnan is a researcher whose work sits at the intersection of computer vision, robotics, and semantic scene understanding. His early contributions focused on visual exploration algorithms for indoor environments, where he pioneered the use of semantic cues to construct hybrid maps that combine topological and semantic representations—a foundational approach for autonomous navigation. His most cited paper (26 citations) introduced a vision-based exploration algorithm that builds image-based hybrid maps, using semantic constructs as nodes to guide robotic movement through unknown spaces. More recently, Krishnan has advanced the field of zero-shot instance segmentation with his work on SupeRGB-D, addressing the critical challenge of detecting and segmenting objects in cluttered indoor environments without requiring extensive manual annotation. This work, which has garnered 11 citations since 2023, is particularly impactful for robots navigating spaces with many small, overlapping objects where traditional 3D sensing falls short. By leveraging deep learning in a zero-shot paradigm, Krishnan’s approach reduces the need for labor-intensive labeled datasets, making it more practical for real-world deployment. His research demonstrates a clear trajectory from early exploration algorithms to cutting-edge segmentation methods, consistently pushing the boundaries of how robots perceive and interact with complex, unstructured environments.