Hema Swetha Koppula

About

Hema Swetha Koppula is a prominent researcher at the intersection of computer vision, robotics, and machine learning, with a particular focus on enabling robots to understand and anticipate human behavior in real-world environments. Her work has made foundational contributions to three interconnected areas: semantic understanding of 3D environments, activity recognition from RGB-D data, and anticipatory robotic response. Koppula's most celebrated work — "Learning Human Activities and Object Affordances from RGB-D Videos" (2013, 699 citations) — demonstrated how robots could extract rich, structured descriptions of human sub-activities by jointly modeling people and the objects they interact with. This was extended powerfully in her highly cited work on anticipating human activities (2015, 613 citations), which showed that robots could predict what a person will do next and respond proactively — a breakthrough for assistive robotics. Her earlier contributions on semantic labeling of 3D point clouds (2011, 331 citations) helped lay the perceptual groundwork for indoor robotic navigation. Beyond individual papers, Koppula contributed to RoboBrain, an ambitious large-scale knowledge engine designed to help robots share learned representations across tasks and modalities. With over 2,300 cumulative citations, her body of work continues to shape how intelligent systems perceive and collaborate with humans.

Research Focus

Key Achievements

Top Papers

1. 1
   Learning human activities and object affordances from RGB-D videos

   699 citations · 2013
2. 2
   Anticipating Human Activities Using Object Affordances for Reactive Robotic Response

   613 citations · 2015
3. 3
   Semantic Labeling of 3D Point Clouds for Indoor Scenes

   331 citations · 2011
4. 4
   Contextually guided semantic labeling and search for three-dimensional point clouds

   182 citations · 2012
5. 5
   Anticipating Human Activities using Object Affordances for Reactive Robotic Response

   130 citations · 2013
6. 6
   RoboBrain: Large-Scale Knowledge Engine for Robots

   110 citations · 2014
7. 7
   Anticipating human activities for reactive robotic response

   102 citations · 2013
8. 8
   Anticipatory Planning for Human-Robot Teams

   69 citations · 2015
9. 9
   Contextually Guided Semantic Labeling and Search for 3D Point Clouds

   58 citations · 2011
10. 10
    Physically Grounded Spatio-temporal Object Affordances

    53 citations · 2014

Key Collaborators

Contact & Links