Shi Detector
Shi detectors, a broad term encompassing various object detection systems, aim to accurately identify and locate objects within diverse data types, from images and videos to event streams and even complex physical phenomena like neutrino interactions. Current research emphasizes improving detector robustness and efficiency through techniques like transformer-based architectures, generative models for data augmentation, and novel loss functions to address challenges such as class imbalance, domain shifts, and adversarial attacks. These advancements have significant implications across numerous fields, including autonomous driving, remote sensing, high-energy physics, and cybersecurity, by enhancing the reliability and accuracy of object detection in real-world applications.
Papers
Self-improving object detection via disagreement reconciliation
Gianluca Scarpellini, Stefano Rosa, Pietro Morerio, Lorenzo Natale, Alessio Del Bue
Assessing Domain Gap for Continual Domain Adaptation in Object Detection
Anh-Dzung Doan, Bach Long Nguyen, Surabhi Gupta, Ian Reid, Markus Wagner, Tat-Jun Chin