Visual Object Tracking and Segmentation challenge VOTS2026 is a continuation of the VOTS2023 and VOTS2024 challenges, which no longer distinct between single- and multi-target tracking nor between short- and long-term tracking. It requires tracking one or more targets simultaneously by segmentation over long or short sequences, while the targets may disappear during tracking and reappear later in the video.
Four challenges are organized
VOTS2026 Challenge Continuation of the VOTS2025 challenge with the task to track one or more general targets over short-term or long-term sequences by segmentation. The dataset is refreshed with new challenging sequences.
VOTSt2026 Challenge Continuation of the VOTSt2025 transformative object tracking challenge, which considers general objects undergoing a topological transformation, such as vegetables cut into pieces, machines disassembled, etc.
VOTSp2026 ChallengeA new point-tracking challenge, which considers tracking several manually defined points. Challenging sequences have been selected to construct a new test set for his challenge.
VOTSr2026 ChallengeA new referring tracking challenge, with the task to segment objects in video based on descriptions.
Winners
VOTS2026 challenge winner “VLM-SAFEDAM -- VLM-DAM4SAM3 with anti-overlap mechanism” Dayu Wu, Yucheng Bai, Lv Wu, Han Fang, Shanghai Jiao Tong University
VOTSt2026 challenge winner “MUSMEM -- Associating Visual Objects Through Transformers” Elham Soltani Kazemi, Gani Rahmon, Imad Eddine Toubal, Juan Mogollon, Kannappan Palaniappan, University of Missouri
VOTSp2026 challenge winner “AllTrackerPlus -- Memory-Efficient AllTracker Wrapper” Zhenglin Du, Zhengyang Li, Yi Wen, Xidian University
VOTSr2026 challenge winner “ReflexTrack -- Feedback-Driven Referring Video Segmentation” Yuanjia Li, Tianyang Xu, ZhangyongTang, He Wang, Xue-Feng Zhu, Xiao-Jun Wu, Josef Kittler; Jiangnan University & University of Surrey
The VOTS2026 workshop was held on September 9, 2026, at ECCV2026 in Copenhagen, Denmark. The workshop featured presentations of the challenge results, invited talks, and discussions on the future of visual object tracking and segmentation research. It was attended by around 70 participants.
Lost items: After the workshop the organizers found two pieces of forgotten tech in the room: a MacBook Pro laptop and a ReMarkable tablet. If you are the owner of these items, please contact the organizers to find out how to retrieve them.
