Patricio Loncomilla
Papers
22
Total Citations
315
H-Index
9
About
Patricio Loncomilla is a Chilean researcher whose work sits at the intersection of computer vision, machine learning, and robotics, with a particular focus on object recognition and robot perception. He has made sustained contributions to the development of visual recognition systems for robotic applications, pioneering the use of local invariant features — most notably SIFT-based methods — for tasks ranging from object manipulation to gaze direction estimation in autonomous robots. His 2016 survey on object recognition using local invariant features, which has garnered over 112 citations, stands as a landmark reference in the field, while his 2018 deep learning survey reflects his ability to track and synthesize emerging paradigm shifts in robot vision. Loncomilla's research extends into applied domains, including RoboCup robot soccer, where he developed tools for automated refereeing and wide-baseline object matching under real-world constraints. His Bayesian methodology for indirect object search and comparative studies of domestic robot manipulation demonstrate a commitment to practical, deployable solutions. With nearly 250 cumulative citations across a decade of work, Loncomilla has established himself as an influential voice bridging classical computer vision techniques and modern deep learning approaches in robotics.
Research Focus
Key Achievements
Top Papers
- 1
- 2A Survey on Deep Learning Methods for Robot Vision50 citations · 2018
- 3Improving SIFT-Based Object Recognition for Robot Applications19 citations · 2005
- 4
- 5Gaze Direction Determination of Opponents and Teammates in Robot Soccer15 citations · 2006
- 6A Bayesian based Methodology for Indirect Object Search14 citations · 2017
- 7
- 8
- 9Applications of Deep Learning in Robot Vision9 citations · 2020
- 10An Automated Refereeing and Analysis Tool for the Four-Legged League9 citations · 2007