The presentation will cover recent classes of methods in General Object Tracking: Common Neural Nets, Correlation Filters and Siamese Trackers. We will focus on three main papers:
- One with an older solution coming back in the spotlight with the new and larger training datasets
- Two having current state-of-the-art on all benchmarks (older and newer)
1. "MDNet: Learning Multi-Domain Convolutional Neural Networks for Visual Tracking, H. Nam, B. Han, CVPR'16, http://cvlab.postech.ac.kr/research/mdnet/
2. "ATOM: Accurate Tracking by Overlap Maximization", M. Danelljan et al.
, CVPR'19, https://github.com/visionml/pytracking
3. "SiamRPN++: Evolution of Siamese Visual Tracking with Very Deep Networks", B. Li, CVPR'19, https://bo-li.info/SiamRPN++/
Presentation will be held in Romanian.
Speaker: Elena Burceanu
Location: PR002 - in the new building of the Faculty of Automatic Control and Computers (PRECIS building - Iuliu Maniu entrance), UPB.
Claim the event and start manage its content.I am the organizer