Computer Vision at Scale: Multi-Camera Tracking, Calibration, and Event Detection for Checkout-Free Retail

Recording

Watch on YouTube, or see the CVPR 2026 virtual tutorial page.

About

Checkout-free retail represents one of the most challenging real-world computer vision deployments, requiring reliable performance across hundreds of sites and millions of interactions. This tutorial bridges academic research and production deployment by framing checkout-free retail as a canonical large-scale multi-camera vision systems problem.

The tutorial focuses on three foundational pillars—presented as generalizable computer vision problems rather than application-specific solutions. Each component is discussed in terms of its underlying formulations, scalability constraints, failure modes, and design tradeoffs that transfer directly to autonomous driving, smart spaces, sports analytics, warehouses, and urban sensing.

A central theme is how infrastructure constraints—including limited bandwidth, latency requirements, camera reliability, and edge computing budgets—fundamentally shape algorithmic and architectural decisions. Attendees will learn how classical and modern deep learning techniques are adapted for continuous online operation, partial observability, and 99.9%+ system reliability.

Key Learning Objectives

• Master automatic multi-camera calibration under scale, drift, and partial observability
• Understand real-time tracking under asynchrony and missing data
• Select cameras and compute architectures for production multi-camera systems
• Recognize how infrastructure constraints drive architectural decisions
• Design systems for reliability at scale

Live Demonstration

Attendees will interact with a live multi-camera perception system demonstrating online calibration, global multi-object tracking, and reliability under realistic bandwidth, latency, and hardware constraints. The system operates fully offline using local compute and networking, with pre-recorded fallback visualizations for all scenarios. Each component (calibration, tracking, reliability handling) can be demonstrated independently.

Broader Impact

While checkout-free retail serves as the motivating application, the tutorial deliberately abstracts away domain-specific details to focus on generalizable formulations applicable to autonomous vehicles, warehouses, smart cities, and sports analytics—any application requiring automatic multi-camera calibration, real-time edge analysis, robust event detection with limited bandwidth, and scale deployment. The systems discussed have contributed to large-scale commercial deployments recognized through industry innovation awards and government-recognized R&D programs.

Schedule

Organizers

Resources

Demo Code

The live demo source is on GitHub: checkout-free-tutorial-cvpr26/demo-code.

Slides

The presentation slides will be available 🔜

References

1. Sarlin et al., "SuperGlue: Learning Feature Matching with Graph Neural Networks," CVPR 2020
2. Sun et al., "LoFTR: Detector-Free Local Feature Matching with Transformers," CVPR 2021
3. Wojke et al., "DeepSORT: Simple Online and Realtime Tracking," 2017
4. Zhang et al., "ByteTrack: Multi-Object Tracking by Associating Every Detection Box," 2021
5. Schönberger & Frahm, "Structure-from-Motion Revisited," CVPR 2016
6. Bandara et al., "AdaMAE: Adaptive Masking for Efficient Spatiotemporal Learning," CVPR 2023 (co-authored by M. Agarwal)
7. Agrawal et al., "Training Data Acquisition for Automated Checkout," US Patent US20220327511A1, 2022
8. Bernardin & Stiefelhagen, "Evaluating Multiple Object Tracking Performance: The CLEAR MOT Metrics," EURASIP 2008
9. Carreira & Zisserman, "Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset," CVPR 2017
10. Lane et al., "DeepX: A Software Accelerator for Low-Power Deep Learning Inference on Mobile Devices," IPSN 2016
11. Hartley & Zisserman, "Multiple View Geometry in Computer Vision (2nd ed.)," Cambridge University Press 2004
12. Longuet-Higgins, "A Computer Algorithm for Reconstructing a Scene from Two Projections," Nature 1981
13. Lowe, "Distinctive Image Features from Scale-Invariant Keypoints (SIFT)," IJCV 2004
14. Rublee et al., "ORB: An Efficient Alternative to SIFT or SURF," ICCV 2011
15. DeTone et al., "SuperPoint: Self-Supervised Interest Point Detection and Description," CVPR Workshops 2018
16. Lindenberger et al., "LightGlue: Local Feature Matching at Light Speed," ICCV 2023
17. Pan et al., "Global Structure-from-Motion Revisited (GLOMAP)," ECCV 2024
18. Bewley et al., "SORT: Simple Online and Realtime Tracking," ICIP 2016
19. Brasó & Leal-Taixé, "Learning a Neural Solver for Multiple Object Tracking," CVPR 2020
20. Cetintas et al., "Unifying Short and Long-Term Tracking with Graph Hierarchies (SUSHI)," CVPR 2023
21. Cokbas et al., "FRIDA: Fisheye Re-Identification Dataset with Annotations," ICIP 2022
22. Collins, "A Space-Sweep Approach to True Multi-Image Matching," CVPR 1996
23. Khan & Shah, "Tracking Multiple Occluding People by Localizing on Multiple Scene Planes," IEEE TPAMI 2009
24. Herzog et al., "Spatial-Temporal Multi-Cuts for Online Multiple-Camera Tracking," arXiv 2024
25. Lin et al., "Camera Link Model for Multi-Camera Tracking," arXiv 2024
26. Hsu et al., "Multi-Camera Tracking with Temporal Anchors," CVPRW 2020
27. Mo et al., "BEV-SUSHI: Multi-Target Multi-Camera 3D Detection and Tracking in Bird's-Eye View," arXiv 2024
28. Nguyen et al., "LMGP: Lifted Multiscale Graph Pyramid Network," CVPR 2022
29. Tang et al., "Monocular Multi-Human Detection Using Grabbed Segments," ICCV 2015
30. H. Zhang et al., "Multi-Camera Tracking with General-Purpose Cameras," CVPR 2008
31. Ye et al., "Deep Learning for Person Re-Identification: A Survey and Outlook," IEEE TPAMI 2021
32. Hermans et al., "In Defense of the Triplet Loss for Person Re-Identification," 2017
33. Zhou et al., "Omni-Scale Feature Learning for Person Re-Identification (OSNet)," ICCV 2019
34. Luo et al., "Bag of Tricks and a Strong Baseline for Deep Person Re-Identification," CVPR Workshops 2019
35. S. He et al., "TransReID: Transformer-Based Object Re-Identification," ICCV 2021
36. Li et al., "CLIP-ReID: Exploiting Vision-Language Model for Image Re-Identification," AAAI 2023
37. T. Chen et al., "A Simple Framework for Contrastive Learning of Visual Representations (SimCLR)," ICML 2020
38. K. He et al., "Momentum Contrast for Unsupervised Visual Representation Learning (MoCo)," CVPR 2020
39. W. Chen et al., "Beyond Appearance: A Semantic Controllable Self-Supervised Learning Framework (SOLIDER)," CVPR 2023