Thomas J. Whelan

About

Thomas J. Whelan is a leading researcher in robotics and computer vision, whose work has fundamentally advanced dense 3D mapping, visual SLAM, and scene understanding. His core contributions lie in developing robust, real-time systems for RGB-D perception, enabling robots to build detailed, persistent maps of their environments. Whelan is perhaps best known for his work on the Kintinuous algorithm, which extended KinectFusion for spatially extended mapping, and for pioneering the Replica dataset—a highly photo-realistic 3D indoor scene dataset with over 384 citations that has become a benchmark for embodied AI research. His papers on robust visual odometry and lifelong object segmentation from change detection (with over 321 and 79 citations, respectively) have set standards for reliable tracking and autonomous learning in dynamic spaces. Whelan’s research also spans humanoid locomotion over uneven terrain and efficient point cloud simplification, demonstrating a rare ability to bridge perception, mapping, and real-world robotic control. With several papers cited hundreds of times, his work has directly shaped modern dense SLAM and continues to influence how robots perceive and interact with complex indoor environments.

Research Focus

Key Achievements

Top Papers

1. 1
   The Replica Dataset: A Digital Replica of Indoor Spaces

   384 citations · 2019
2. 2
   Robust real-time visual odometry for dense RGB-D mapping

   321 citations · 2013
3. 3
   Toward lifelong object segmentation from change detection in dense RGB-D maps

   79 citations · 2013
4. 4
   Continuous humanoid locomotion over uneven terrain using stereo fusion

   75 citations · 2015
5. 5
   Robust Tracking for Real-Time Dense RGB-D Mapping with Kintinuous

   47 citations · 2012
6. 6
   Robotics: Science and Systems XI

   27 citations · 2015
7. 7
   Planar simplification and texturing of dense point cloud maps

   20 citations · 2013
8. 8
   Robust Tracking for Real-Time Dense RGB-D Mapping with Kintinuous : Computer Science and Artificial Intelligence LaboratoryTechnical Report MIT-CSAIL-TR-2012-031

   9 citations · 2012
9. 9
   Line Point Registration: A Technique for Enhancing Robot Localization in a Soccer Environment

   4 citations · 2012
10. 10
    3D Mapping, Localisation and Object Retrieval using Low Cost Robotic Platforms: A Robotic Search Engine for the Real-World

    3 citations · 2014

Key Collaborators

Contact & Links