Multi Object Tracking | Google-Colab

Collection of papers, code, datasets, and other resources for multi object tracking | Google colab

Joint-Detection

• CenterTrack - Tracking Objects as Points [ax2004] [pdf] [notes] [code] [pytorch]

• DEFT: Detection Embeddings for Tracking [ax2102] [arxiv] [pdf] [notes] [code] [pytorch]

Identity Embedding

• MOTS Multi-Object Tracking and Segmentation [cvpr19] [pdf] [notes] [code] [project/data]

• Towards Real-Time Multi-Object Tracking [ax1909] [arxiv] [pdf] [notes] [code]

  • Towards Real-Time Multi-Object Tracking - Vehicle Tracking [Modified version]

• A Simple Baseline for Multi-Object Tracking [ax2004] [pdf] [notes] [code]

• Integrated Object Detection and Tracking with Tracklet-Conditioned Detection [ax1811] [pdf] [notes]

Association

• Deep Affinity Network for Multiple Object Tracking [ax1810/tpami19] [pdf] [notes] [code] [pytorch]

Deep Learning

• Online Multi-Object Tracking Using CNN-based Single Object Tracker with Spatial-Temporal Attention Mechanism [ax1708/iccv17] [pdf] [arxiv] [notes]

• Online multi-object tracking with dual matching attention networks [ax1902/eccv18] [pdf] [arxiv] [notes] [code]

• FAMNet Joint Learning of Feature, Affinity and Multi-Dimensional Assignment for Online Multiple Object Tracking [iccv19] [pdf] [notes]

• Exploit the Connectivity: Multi-Object Tracking with TrackletNet [ax1811/mm19] [pdf] [notes]

• Tracking without bells and whistles [ax1903/iccv19] [pdf] [notes] [code] [pytorch]

  • Train object detector (Faster-RCNN with FPN)

RNN

• Tracking The Untrackable: Learning To Track Multiple Cues with Long-Term Dependencies [ax1704/iccv17] [Stanford] [pdf] [notes] [arxiv] [project],
• Multi-object Tracking with Neural Gating Using Bilinear LSTM [eccv18] [pdf] [notes]
• Eliminating Exposure Bias and Metric Mismatch in Multiple Object Tracking [cvpr19] [pdf] [notes] [code]

Unsupervised Learning

• Unsupervised Person Re-identification by Deep Learning Tracklet Association [ax1809/eccv18] [pdf] [notes]
• Tracking by Animation: Unsupervised Learning of Multi-Object Attentive Trackers [ax1809/cvpr19] [pdf] [arxiv] [notes] [code]
• Simple Unsupervised Multi-Object Tracking [ax2006] [pdf] [notes]

Reinforcement Learning

• Learning to Track: Online Multi-object Tracking by Decision Making [iccv15] [Stanford] [pdf] [notes] [code (matlab)] [project]
• Collaborative Deep Reinforcement Learning for Multi-Object Tracking [eccv18] [pdf] [notes]

Network Flow

• Near-Online Multi-target Tracking with Aggregated Local Flow Descriptor [iccv15] [NEC Labs] [pdf] [author] [notes]
• Deep Network Flow for Multi-Object Tracking [cvpr17] [NEC Labs] [pdf] [supplementary] [notes]
• Learning a Neural Solver for Multiple Object Tracking [ax1912/cvpr20] [pdf] [notes] [code]

Graph Optimization

• A Multi-cut Formulation for Joint Segmentation and Tracking of Multiple Objects [ax1607] [highest MT on MOT2015] [University of Freiburg, Germany] [pdf] [arxiv] [author] [notes]

Baseline

• Simple Online and Realtime Tracking [icip16] [pdf] [notes] [code]
• Simple Online and Realtime Tracking with a Deep Association Metric [icip17] [arxiv] [code] [code2]

• High-Speed Tracking-by-Detection Without Using Image Information [avss17] [pdf] [notes] [code]

Metrics

• HOTA A Higher Order Metric for Evaluating Multi-object Tracking [ijcv20/08] [pdf] [notes] [code]

Multi Object Tracking Datasets

• IDOT
• UA-DETRAC Benchmark Suite
• GRAM Road-Traffic Monitoring
• Stanford Drone Dataset
• Ko-PER Intersection Dataset
• TRANCOS
• Urban Tracker
• DARPA VIVID / PETS 2005 [Non stationary camera]
• KIT-AKS [No ground truth]
• CBCL StreetScenes Challenge Framework [No top down viewpoint]
• MOT 2015 [mostly street level viewpoint]
• MOT 2016 [mostly street level viewpoint]
• MOT 2017 [mostly street level viewpoint]
• MOT 2020 [mostly top down viewpoint]
• MOTS: Multi-Object Tracking and Segmentation [MOT and KITTI]
• CVPR 2019 [mostly street level viewpoint]
• PETS 2009 [No vehicles]
• PETS 2017 [Low density] [mostly pedestrians]
• DukeMTMC [multi camera] [static background] [pedestrians] [above-street level viewpoint] [website not working]
• KITTI Tracking Dataset [No top down viewpoint] [non stationary camera]
• The WILDTRACK Seven-Camera HD Dataset [pedestrian detection and tracking]
• 3D Traffic Scene Understanding from Movable Platforms [intersection traffic] [stereo setup] [moving camera]
• LOST : Longterm Observation of Scenes with Tracks [top down and street level viewpoint] [no ground truth]
• JTA [top down and street level viewpoint] [synthetic/GTA 5] [pedestrian] [3D annotations]
• PathTrack: Fast Trajectory Annotation with Path Supervision [top down and street level viewpoint] [iccv17] [pedestrian]
• CityFlow [pole mounted] [intersections] [vehicles] [re-id] [cvpr19]
• UAVDT - The Unmanned Aerial Vehicle Benchmark: Object Detection and Tracking [uav] [intersections/highways] [vehicles] [eccv18]
• JackRabbot Dataset [RGBD] [head-on][indoor/outdoor][stanford]
• TAO: A Large-Scale Benchmark for Tracking Any Object [eccv20] [code]

Acknowledgements

• Part of README.md is borrowed from the Deep-Learning-for-Tracking-and-Detection check out this amazing github repo for other detection & tracking resources.