Nicolas Gauthier

Institute for Infocomm Research, Agency for Science, Technology and Research

Papers

Total Citations

H-Index

About

Nicolas Gauthier is a leading researcher in active vision and robotic perception, whose work addresses fundamental challenges in how robots efficiently interact with dynamic, open-world environments. His primary research areas include active object detection, view planning, and human-robot collaboration, with a focus on reducing computational and data costs while improving robustness. Gauthier’s major contributions center on adaptive view planning—a technique that allows robots to intelligently select camera angles to maximize detection accuracy and minimize uncertainty. His 2021 paper “Towards Efficient Multiview Object Detection with Adaptive Action Prediction” (10 citations) introduces a reinforcement learning framework that predicts optimal actions, significantly boosting efficiency over traditional methods. In “Enhancing Multi-Step Action Prediction for Active Object Detection” (8 citations), he extends this to multi-step planning, further improving performance. His 2020 work “Active Image Sampling on Canonical Views for Novel Object Detection” (7 citations) leverages canonical view models to reduce annotation burdens, a key step toward scalable learning. Additionally, Gauthier’s “Task-Oriented Multi-Modal Question Answering For Collaborative Applications” (2020, 4 citations) pioneers a new QA dataset for human-robot collaboration, showcasing his commitment to practical, task-driven robotics. With a growing citation impact, Gauthier’s research is shaping the future of efficient, adaptive robotic vision.