Thomas A. Ciarfuglia
University of Perugia, Sapienza University of Rome, San Raffaele University of Rome
Papers
17
Total Citations
433
H-Index
9
About
Thomas A. Ciarfuglia is a computer vision and robotics researcher whose work spans autonomous navigation, depth estimation, and precision agriculture. He is best known for his contributions to monocular depth estimation, where his 2016 paper on fast robust depth estimation using fully convolutional networks for obstacle detection garnered 105 citations — establishing him as a key voice in perception systems for high-speed autonomous platforms. His follow-up work on domain-independent learning-based depth estimation (69 citations) further pushed the field toward more generalizable, geometry-independent solutions suitable for real-world deployment. Beyond depth perception, Ciarfuglia has made notable contributions to visual odometry evaluation (49 citations), loop closure detection using CNN features and covisibility graphs (48 citations), and natural language video description for human-robot interaction (35 citations), reflecting a broad commitment to robust robotic perception and communication. More recently, his research has turned toward precision agriculture, with impactful work on weakly supervised detection and tracking of table grapes (39 citations) and the AgriSORT real-time multi-object tracking framework (20 citations). With over 400 cumulative citations and contributions spanning nearly a decade, Ciarfuglia represents a versatile researcher bridging foundational computer vision with applied agricultural robotics.
Research Focus
Key Achievements
Top Papers
- 1
- 2Toward Domain Independence for Learning-Based Monocular Depth Estimation69 citations · 2017
- 3Evaluation of non-geometric methods for visual odometry49 citations · 2014
- 4
- 5
- 6
- 7A transfer learning approach for multi-cue semantic place recognition26 citations · 2013
- 8
- 9Pseudo-label Generation for Agricultural Robotics Applications9 citations · 2022
- 10