Dario Mantegazza

University of Applied Sciences and Arts of Southern Switzerland

Papers

Total Citations

H-Index

About

Dario Mantegazza is a robotics and computer vision researcher whose work centers on visual anomaly detection systems for mobile robots — a critical challenge in ensuring safe and reliable autonomous operation. His research addresses the fundamental problem of enabling robots to identify hazardous or unexpected situations from visual input, even when training data is scarce or imbalanced. Mantegazza's most influential contribution, "An Outlier Exposure Approach to Improve Visual Anomaly Detection Performance for Mobile Robots" (2022, 15 citations), introduced a practical methodology for leveraging limited examples of anomalous data to significantly boost detection performance beyond what standard unsupervised models achieve. This work directly tackles real-world robotics constraints where anomaly examples are rare but high-stakes. Complementing this, his development of the **Hazards&Robots** dataset (2023, 7 citations) — comprising over 324,000 RGB frames across 20 anomaly classes — provides the research community with a standardized benchmark that was previously lacking in this domain. His parallel work on sensing anomalies as potential hazards further consolidates a coherent research agenda around safety-critical perception. With growing citation momentum across multiple publications, Mantegazza is establishing himself as a focused contributor to the intersection of machine learning and dependable robotics.