Detail of the student project

List
Topic:Rozpoznávání objektů v simulovaném prostředí
Department:Katedra kybernetiky
Supervisor:doc. Ing. Tomáš Pajdla, Ph.D.
Announce as:Bakalářská práce, Semestrální projekt
Description:Collect a set of objects that can and can't be recognized by a state of the art recognition/semantic segmentation method. Investigate and describe how the recognition behaves in a state of the art simulation. Suggest and demonstrate a technique for detecting unkonw classes.
Bibliography:1) Bayesian Semantic Instance Segmentation in Open Set World https://arxiv.org/pdf/1806.00911.pdf
2) The Importance of Metric Learning for Robotic Vision: Open Set Recognition and Active Learning https://arxiv.org/pdf/1902.10363.pdf
3) Recent Advances in Open Set Recognition: A Survey https://arxiv.org/pdf/1811.08581.pdf
4) YOLACT Real-time Instance Segmentation. https://openaccess.thecvf.com/content_ICCV_2019/papers/Bolya_YOLACT_Real-Time_Instance_Segmentation_ICCV_2019_paper.pdf
Responsible person: Petr Pošík