Nicolai Petkov
Papers
17
Total Citations
396
H-Index
12
About
Nicolai Petkov is a researcher whose work spans computer vision, robotics, and biologically inspired computational models, with particular emphasis on visual place recognition, semantic mapping, and object detection. His most influential contributions center on applying convolutional neural networks to appearance-invariant place recognition — a critical challenge in autonomous driving and mobile robotics — with his 2017 paper accumulating 78 citations and a related 2015 study earning 52, reflecting sustained community interest in this direction. Petkov has also made notable strides in modeling the primate visual system, proposing the trainable S-COSFIRE hierarchical object recognition framework and extending COSFIRE-based filters to crack delineation and curved line detection. His robotics research is notably interdisciplinary, encompassing semantic indoor mapping through the ViMantic architecture, human action recognition from skeletal data, and even odor recognition for robotic applications. His involvement in the TrimBot2020 project demonstrates a commitment to real-world deployment of intelligent outdoor robots. Collectively, his portfolio reflects a researcher who bridges neuroscience-inspired modeling with practical autonomous systems, earning over 300 cumulative citations and contributing meaningfully to both foundational theory and applied robotics.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5Learning skeleton representations for human action recognition29 citations · 2018
- 6
- 7
- 8
- 9TB-Places: A Data Set for Visual Place Recognition in Garden Environments21 citations · 2019
- 10TrimBot2020: an outdoor robot for automatic gardening20 citations · 2018