7 datasets found

MOT16 is a benchmark for multi-object tracking, consisting of 16 video sequences with 1000 frames each.

The Visual Tracking Benchmark (VTB) is a collection of 50 videos.

Urban Tracker dataset for multiple object tracking in urban mixed traffic

The CLEAR MOT metrics are used to evaluate the performance of multi-object tracking algorithms.

Unsupervised multiple-object tracking with a dynamical variational autoencoder

The proposed architecture is described in Fig. 3 which is based on TransTrack [21] with several improvements to reduce the computational complexity and model size.

Many Multi-Object Tracking (MOT) approaches exploit motion information to associate all the detected objects across frames. However, many methods that rely on filtering-based...