V7 allows you to label, train, and deploy vision AI models and handle your team's ML-Ops stack.
try It for Free

PPE Detection

Load videos or images of worker sites to detect the usage of PPE. Label images using automated annotation to specify what PPE looks like in your environment, and easily train a deep neural network in one click. Take a look below at what teams on V7 built for PPE detection.

Hover over the image below to see what the AI identifies

worker PPE detection computer vision AIworker PPE detection computer vision AI

See it in action

Detecting PPE in video with machine learning

According to NIOSH, over 2,000 work-related injuries occur every day in the US which could be prevented through the use of PPE. Deep learning has enabled the detection of protective equipment across multiple domains, as long as training data is varied and representative enough. To leverage V7 for PPE detection, start by uploading your domain's video footage or images displaying protective equipment such as hardhats, safety glasses, or face masks. Label them using V7's annotation tools, and then train a model to adapt to a new camera deployment or set of equipment.

What can be identified?

PPE comes in many forms, items that can be reliably spotted with machine learning are:

  • Hardhats
  • Face masks
  • High visibility jackets
  • Welding goggles and welding masks
  • Gloves
  • Joint protectors
  • Ear protectors
  • Safety glasses

How much training data is needed for a reliable PPE detection?

The best ML models start to learn objects across domains after they've seen 1,000 examples, but you can start training with as little as 30 on V7. What's important is that you fine-tune your model upon new deployments. This means uploading a small sample of data from the new factory, construction site, or laboratory where the PPE detection is happening, so that the AI model adapts to what the camera quality, background, and subjects look like. There are no limitations when it comes to lighting or angles to worry about, however make sure your object's sizes are at least 1% of the image's width. If they aren't, consider labelling a taxonomy level above that (for example rather than "safety glasses" try "face with safety glasses".

View additional examples

Click to enlarge the images below

No items found.

Case Study

No items found.

From the Blog

ML-Ops defines your accuracy. We help you grow it to 99.9%.

AI performance shouldn't stop at the first training session.
Sync your product with a dataset and continually improve your AI performance by turning its output into new ground truth. Continual learning will allow your models to surpass 95% accuracy barriers and reach 99% and above as more data from your use case or product is supervised by humans, and learnt by your AI.

v7 neurons product
Your product
accurate labeling with v7 darwin
Fast labeling on V7 Darwin
image dataset
Newly collected images or video
model training and pruning
Training & active learning

Ready to get started?

Schedule a demo with our team or discuss your project.

Dataset Management

AutoML model training to solve visual tasks or auto-label your datasets, and a scalable inference engine to launch your project.