Papers
8
Total Citations
295
H-Index
7
About
Quentin Bateux is a robotics researcher specializing in visual servoing, deep learning-based robot control, and compliant manipulation. His work has fundamentally advanced the field of visual servoing — the use of visual feedback to guide robot motion — by bridging classical control theory with modern machine learning techniques. Bateux's most influential contribution, "Training Deep Neural Networks for Visual Servoing" (2018, 157 citations), introduced a convolutional neural network approach capable of performing high-precision, real-time 6 DOF positioning tasks with remarkable robustness to occlusions and lighting variation. This work, building on his earlier "Visual Servoing from Deep Neural Networks" (2017), demonstrated that deep learning could replace hand-engineered feature extraction in complex robotic control pipelines. Complementing this, his series of papers on histogram-based and photometric visual servoing expanded the convergence domain of direct methods, addressing a longstanding limitation in the field. More recently, Bateux has extended his expertise into compliant manipulation and contact-rich assembly, exploring generalized robot assembly strategies and vision-based force estimation without dedicated force sensors. His cumulative body of work, spanning foundational methodology to practical robotics applications, has garnered nearly 300 citations, establishing him as a meaningful contributor to intelligent robot perception and control.
Research Focus
Key Achievements
Top Papers
- 1Training Deep Neural Networks for Visual Servoing157 citations · 2018
- 2Histograms-Based Visual Servoing58 citations · 2016
- 3Visual Servoing from Deep Neural Networks30 citations · 2017
- 4Direct visual servoing based on multiple intensity histograms19 citations · 2015
- 5
- 6Particle filter-based direct visual servoing10 citations · 2016
- 7
- 8Going further with direct visual servoing4 citations · 2018