Reference : ARTHuS: Adaptive Real-Time Human Segmentation in Sports through Online Distillation
Scientific congresses and symposiums : Paper published in a journal
Engineering, computing & technology : Electrical & electronics engineering
http://hdl.handle.net/2268/234413
ARTHuS: Adaptive Real-Time Human Segmentation in Sports through Online Distillation
English
Cioppa, Anthony* mailto [Université de Liège - ULiège > Dép. d'électric., électron. et informat. (Inst.Montefiore) > Télécommunications >]
Deliège, Adrien* mailto [Université de Liège - ULiège > Dép. d'électric., électron. et informat. (Inst.Montefiore) > Télécommunications >]
Istasse, Maxime mailto [UCLouvain > > > >]
De Vleeschouwer, Christophe mailto [UCLouvain > > > >]
Van Droogenbroeck, Marc mailto [Université de Liège - ULiège > Dép. d'électric., électron. et informat. (Inst.Montefiore) > Télécommunications >]
* These authors have contributed equally to this work.
In press
IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) Proceedings
Yes
International
IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) - CVSports
du 16 juin 2019 au 20 juin 2019
IEEE
Long Beach
CA
[en] semantic segmentation ; human segmentation ; real-time ; online distillation ; deep learning ; artificial intelligence ; computer vision ; soccer ; basketball ; sports ; players ; ARTHuS
[en] Semantic segmentation can be regarded as a useful tool for global scene understanding in many areas, including sports, but has inherent difficulties, such as the need for pixel-wise annotated training data and the absence of well-performing real-time universal algorithms. To alleviate these issues, we sacrifice universality by developing a general method, named ARTHuS, that produces adaptive real-time match-specific networks for human segmentation in sports videos, without requiring any manual annotation. This is done by an online knowledge distillation process, in which a fast student network is trained to mimic the output of an existing slow but effective universal teacher network, while being periodically updated to adjust to the latest play conditions. As a result, ARTHuS allows to build highly effective real-time human segmentation networks that evolve through the match and that sometimes outperform their teacher. The usefulness of producing adaptive match-specific networks and their excellent performances are demonstrated quantitatively and qualitatively for soccer and basketball matches.
Région wallonne : Direction générale des Technologies, de la Recherche et de l'Energie - DGTRE
DeepSport
Researchers ; Professionals
http://hdl.handle.net/2268/234413
Accepted for CVPR 2019 Workshop : "Computer vision in Sports (CVSports)".

File(s) associated to this reference

Fulltext file(s):

FileCommentaryVersionSizeAccess
Open access
paper_camera_ready_ORBI.pdfAuthor preprint9.24 MBView/Open
Open access
supplementary_material_camera_ready.pdfAuthor preprint8.15 MBView/Open

Bookmark and Share SFX Query

All documents in ORBi are protected by a user license.