Search: onr:"swepub:oai:DiVA.org:liu-137882" > Adaptive Decontamin...
Fältnamn | Indikatorer | Metadata |
---|---|---|
000 | 03449naa a2200337 4500 | |
001 | oai:DiVA.org:liu-137882 | |
003 | SwePub | |
008 | 170601s2016 | |||||||||||000 ||eng| | |
024 | 7 | a https://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-1378822 URI |
024 | 7 | a https://doi.org/10.1109/CVPR.2016.1592 DOI |
040 | a (SwePub)liu | |
041 | a engb eng | |
042 | 9 SwePub | |
072 | 7 | a ref2 swepub-contenttype |
072 | 7 | a kon2 swepub-publicationtype |
100 | 1 | a Danelljan, Martin,d 1989-u Linköpings universitet,Datorseende,Tekniska fakulteten4 aut0 (Swepub:liu)marda26 |
245 | 1 0 | a Adaptive Decontamination of the Training Set: A Unified Formulation for Discriminative Visual Tracking |
264 | 1 | b Institute of Electrical and Electronics Engineers (IEEE),c 2016 |
338 | a electronic2 rdacarrier | |
500 | a Funding Agencies|SSF (CUAS); VR (EMC2); VR (ELLIIT); Wallenberg Autonomous Systems Program; NSC; Nvidia | |
520 | a Tracking-by-detection methods have demonstrated competitive performance in recent years. In these approaches, the tracking model heavily relies on the quality of the training set. Due to the limited amount of labeled training data, additional samples need to be extracted and labeled by the tracker itself. This often leads to the inclusion of corrupted training samples, due to occlusions, misalignments and other perturbations. Existing tracking-by-detection methods either ignore this problem, or employ a separate component for managing the training set. We propose a novel generic approach for alleviating the problem of corrupted training samples in tracking-by-detection frameworks. Our approach dynamically manages the training set by estimating the quality of the samples. Contrary to existing approaches, we propose a unified formulation by minimizing a single loss over both the target appearance model and the sample quality weights. The joint formulation enables corrupted samples to be down-weighted while increasing the impact of correct ones. Experiments are performed on three benchmarks: OTB-2015 with 100 videos, VOT-2015 with 60 videos, and Temple-Color with 128 videos. On the OTB-2015, our unified formulation significantly improves the baseline, with a gain of 3.8% in mean overlap precision. Finally, our method achieves state-of-the-art results on all three datasets. | |
650 | 7 | a NATURVETENSKAPx Data- och informationsvetenskapx Datorseende och robotik0 (SwePub)102072 hsv//swe |
650 | 7 | a NATURAL SCIENCESx Computer and Information Sciencesx Computer Vision and Robotics0 (SwePub)102072 hsv//eng |
700 | 1 | a Häger, Gustav,d 1988-u Linköpings universitet,Datorseende,Tekniska fakulteten4 aut0 (Swepub:liu)gusha40 |
700 | 1 | a Khan, Fahad Shahbaz,d 1983-u Linköpings universitet,Datorseende,Tekniska fakulteten4 aut0 (Swepub:liu)fahkh30 |
700 | 1 | a Felsberg, Michael,d 1974-u Linköpings universitet,Datorseende,Tekniska fakulteten4 aut0 (Swepub:liu)micfe03 |
710 | 2 | a Linköpings universitetb Datorseende4 org |
773 | 0 | t 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)d : Institute of Electrical and Electronics Engineers (IEEE)g , s. 1430-1438q <1430-1438z 9781467388511z 9781467388528 |
856 | 4 | u https://liu.diva-portal.org/smash/get/diva2:1104732/FULLTEXT02.pdfx primaryx Raw objecty fulltext:postprint |
856 | 4 8 | u https://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-137882 |
856 | 4 8 | u https://doi.org/10.1109/CVPR.2016.159 |
Kungliga biblioteket hanterar dina personuppgifter i enlighet med EU:s dataskyddsförordning (2018), GDPR. Läs mer om hur det funkar här.
Så här hanterar KB dina uppgifter vid användning av denna tjänst.
Copy and save the link in order to return to this view