Papers
14
Total Citations
188
H-Index
7
About
Karthik Desingh is a robotics researcher whose work sits at the intersection of perception, manipulation, and human-robot interaction. His research addresses some of the most persistent challenges in autonomous robotics: enabling robots to understand and act within complex, cluttered real-world environments. Desingh has made notable contributions to probabilistic scene estimation, developing axiomatic and particle filtering approaches that allow robots to reason about occluded and physically interacting objects during goal-directed manipulation — work that has collectively garnered over 57 citations. His semantic mapping framework, CT-Map, advances simultaneous object detection and 6-DoF pose localization, reflecting his commitment to building robots with rich environmental awareness. Desingh has also pushed the boundaries of articulated object manipulation through efficient nonparametric belief propagation, and explored spatial reasoning for sequential tasks via object-centric neural representations. More recently, his research has embraced natural language understanding, with contributions to 3D visual grounding and end-user directed manipulation learning. Spanning probabilistic inference, deep learning, and human-in-the-loop systems, his body of work charts a coherent path toward robots that are both perceptually capable and genuinely useful to human users.
Research Focus
Key Achievements
Top Papers
- 1Goal-directed robot manipulation through axiomatic scene estimation35 citations · 2017
- 2Semantic Mapping with Simultaneous Object Detection and Localization34 citations · 2018
- 3
- 4Axiomatic particle filtering for goal-directed robotic manipulation22 citations · 2015
- 5LanguageRefer: Spatial-Language Model for 3D Visual Grounding22 citations · 2021
- 6EURECA: Enhanced Understanding of Real Environments via Crowd Assistance15 citations · 2018
- 7SORNet: Spatial Object-Centric Representations for Sequential\n Manipulation13 citations · 2021
- 8Sketching Affordances for Human-in-the-loop Robotic Manipulation Tasks5 citations · 2019
- 9Semantic Mapping with Simultaneous Object Detection and Localization5 citations · 2018
- 10Talk Through It: End User Directed Manipulation Learning4 citations · 2024