You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
An open-source project dedicated to tracking and segmenting any objects in videos, either automatically or interactively. The primary algorithms utilized include the Segment Anything Model (SAM) for key-frame segmentation and Associating Objects with Transformers (AOT) for efficient tracking and propagation purposes.
A complete paper list of Single-Object Visual Tracking Algorithms, Surveys and Benchmarks of recent years. Different from existing paper list, this project doesn't simply category the papers by publishment, but from a tracking challenge-tackling perspective.
This work proposes a feature refined end-to-end tracking framework with a balanced performance using a high-level feature refine tracking framework. The feature refine module enhances the target feature representation power that allows the network to capture salient information to locate the target. The attention module is employed inside the fe…