Kaustubh Kulkarni

Institut national de recherche en sciences et technologies du numérique, Directorate-General for Interpretation

Papers

Total Citations

H-Index

About

Kaustubh Kulkarni is a researcher whose work lies at the intersection of robotics, audiovisual perception, and human-robot interaction. His primary contribution is the creation of the RAVEL (Robot Audiovisual Environment for Learning) dataset, a foundational resource for training robots to process and integrate auditory and visual information in natural settings. The dataset, recorded using the POPEYE robot head equipped with two cameras and four microphones, captures the complexities of real-world indoor environments, including background noise and dynamic scenes. Kulkarni’s most cited work, "RAVEL: an annotated corpus for training robots with audiovisual abilities" (2012), has garnered 35 citations, underscoring its influence in advancing multimodal perception systems. By providing a publicly available, richly annotated corpus, Kulkarni has enabled researchers to develop and benchmark algorithms for tasks such as sound localization, speaker identification, and audiovisual scene understanding. His efforts have laid critical groundwork for creating more perceptive and context-aware robots, bridging the gap between controlled laboratory conditions and the messy, unpredictable realities of everyday human environments.