Behaviour & safetyCCTV Shoplifting Detection
Train a shoplifting detection model: 5,000+ synthetic CCTV frames and 100+ videos with YOLO, pose and VLM labels. Free 400-image sample on Kaggle.
- Size
- 5,000+ frames + 100+ videos
- Format
- YOLO + VLM
Video datasets
Video datasets for computer vision
Datasets that include video sequences or evaluation clips, for temporal models and for testing a frame detector the way it will run: on moving footage.
8 datasets
Behaviour & safety
Behaviour & safetyCCTV Fall & Incident Detection
Synthetic fall detection dataset: 1,000 CCTV images of fallen people plus sitting and kneeling look-alikes, labelled person / fallen_person in YOLO and COCO.
- Size
- 1,000 images
- Format
- YOLO + COCO
Counter-UAS & defenceAirport Drone Threat & Safety
Synthetic airport drone detection dataset: 1,200 YOLO images, small-object focus, adverse weather, bird-vs-drone labels. GDPR compliant.
- Size
- 1,200 images
- Format
- YOLO
Weapons & threatsCCTV Weapon Detection: Rifles vs Umbrellas
Synthetic hard-negative dataset: rifles vs umbrellas to cut weapon-detection false positives. 1,000 YOLO images. Open-source sample available.
- Size
- 1,000 images
- Format
- YOLO
Weapons & threatsWeapon Detection: Handgun vs Bag of Chips
Fix the 'Doritos Effect' in weapon detection AI with handgun vs chip-bag hard negatives. 1,100 YOLO images. Open-source sample available.
- Size
- 1,100 images
- Format
- YOLO
Counter-UAS & defenceMilitary Drone Swarm & Saturation Attack
Synthetic drone swarm dataset for Counter-UAS training. 5-15 drones per frame, 10-50 pixel objects, adverse weather. 1,000+ YOLO-annotated images.
- Size
- 1,000+ images
- Format
- YOLO
Counter-UAS & defenceDrone Civilian / Armed Military Detection
Synthetic drone-perspective dataset distinguishing civilians from armed personnel. 800 YOLO images for aerial surveillance AI training.
- Size
- 800 images
- Format
- YOLO
Access nowSample on request
Behaviour & safetyDriver Drowsiness Detection
Synthetic driver drowsiness dataset with 10,000 images and pixel-perfect eye keypoint annotations. 168 open-source samples. Ideal for fatigue detection AI.
- Size
- 10,000 images
- Format
- YOLO + Keypoints
Most detection datasets are still images. Behaviour classes (shoplifting, fights, falls, loitering) and small fast targets (drones) need sequences: per-frame boxes, tracking IDs, pose keypoints and event start and end times. The datasets below ship video, either as the main product (shoplifting) or as evaluation clips alongside the images.
Custom video is also available: clips rendered into photos of your own camera views, with per-frame boxes, tracking IDs and event times, quoted per project.
Prices
One-time licences, per dataset.
| Images | Research | Commercial | Order |
|---|---|---|---|
| 1,000 | $149 | $590 | On the dataset page |
| 5,000 | Not offered | $2,400 | Order 5,000 |
| 10,000 | Not offered | $4,500 | Order 10,000 |
| 20,000 | Not offered | $8,500 | Order 20,000 |
- Commercial licence only: the research licence covers 1,000 images
- YOLO + COCO labels, fixed train / val split, dataset card
- Hard negatives in the same share as the 1,000-image set
- Masks, pose keypoints or VLM captions where the dataset has them
- Card or invoice; delivered as a download link within five working days
Larger packs are assembled for your order and delivered as a download link within five working days. Prices in USD, excl. VAT, one-time, per dataset.
Read next
Guides for this kind of data.
Guide
15 Video Datasets for Computer Vision, Compared (2026)
UCF101, Kinetics, NTU RGB+D, AVA, UCF-Crime, XD-Violence, UBnormal and more compared by size, classes, camera type and licence, plus what none of them covers.
Guide
CCTV Video Anomaly Detection Datasets: A Practical Guide
How CCTV video anomaly detection datasets are built, which behaviors matter (falls, fights, weapon draws, loitering), and why synthetic CCTV wins.
Guide
Human Activity Recognition Datasets: Benchmarks, Labels and Gaps
What HAR data looks like, how activity clips are labeled, why public benchmarks break in deployment, and how synthetic video covers rare activities.
Questions
Video datasets: common questions
What is a video dataset for computer vision?
A set of video clips with labels over time: per-frame bounding boxes, tracking IDs that follow one object across frames, pose keypoints, or clip-level event labels with start and end times.
Which public video datasets exist for CCTV?
UCF-Crime, RWF-2000, XD-Violence, UBnormal, ShanghaiTech and the fall sets CAUCAFall and GMDCSA-24 are the most used; our guide compares 15 of them on size, camera, labels and licence.
How is a synthetic video dataset priced?
The shoplifting video dataset is priced on request. Datasets that include evaluation clips alongside images are sold as image licences (research or commercial), with the clips included. Custom video on your cameras is quoted per project.
More collections
Other ways into the catalogue.
Show us the alarm your model keeps getting wrong.
Send five frames from your cameras. We reply within one working day with the fastest way to fix it — a catalogue dataset, a stress test or a custom pack.