Skip to main content

Vision AI toolkit

Project description

VisionAI

Documentation for VisionAI toolkit.

Overview

VisionAI provides a set of command line utilities for you to manage different Vision AI scenarios that have been pre-developed and pre-tested. VisionAI focuses on workplace health and safety models - and majority of the models you see here have been developed with that in mind.

These are production-ready model trained from open-source and academic datasets. We are continuously working on new scenarios - and our current scenario repo consists of over 60 scenarios that are listed here. They are developed with the intent of being easy-to-use for business. The framework also supports a whole bunch of custom scenarios.

Install VisionAI

Install VisionAI application through PyPI. There are other options available for install - including a Docker container option. These are detailed in installation section.

$ pip install visionai
---> 100%
Successfully installed visionai

✨ You are all set to use visionai toolkit ✨

Deploy to Azure

Deploy a fully configured and tested solution directly from Azure Marketplace. VisionAI runs computer vision models, most of which run orders of magnitude faster if executed on a GPU machine. Our Azure Marketplace offer VisionAI Community Edition is available through Azure Marketplace here (TODO). The community edition deploys a fully configured Virtual Machine with the recommended hardware and software options. Get more details here.

Deploy to Azure

  • TODO: Point to ARM template that needs to be deployed (using these instructions and here is an example JSON file).

List available Scenarios

VisionAI is organized in terms of scenarios. Consider each scenario as being a business use-case, that is solved by a combination of Machine Learning models and an inference algorithm. For example Warn me when max occupancy of this area exceeds 80 people is a business scenario, where as the People detection is an ML model.

VisionAI supports 60 scenarios currently and more are being added continuously. Our current focus is on Workplace Safety scenarios. Please contact us if a scenario you need is not present in our repo and we will look into it.

$ visionai scenarios list

------------------------------------------------
Privacy Suite
blur-faces
blur-text

Fire safety
early-smoke-and-fire-detection
smoking-and-vaping-detection

Personnel safety
ppe-detection
pfas-system-detection
railings-detection

Suspicious activity
shipping-activity-detection
agressive-behaivior


Compliance Policies
max-occupancy

Equipment
rust-and-corrosion-detection

IR Camera
temperature-monitoring
------------------------------------------------

✨ More scenarios are added regularly ✨

Get details for a Scenario

You can get details about a scenario using visionai scenario details command. Specify the scenario you want additional details for. The details of a scenario include the dataset size, model accuracy metrics,

$ visionai scenario --name early-smoke-and-fire-detection details

------------------------------------------------
Category: Fire safety
Scenario: early-smoke-and-fire-detection
This scenario has been trained on open-source datasets consisting of 126,293 images. The datasets images are primarily outdoors (70%), but do contain a good number of indoor images (30%). There is a ~50-50% mix of day vs night images. You can find more details about this scenario at visionify.ai/early-smoke-and-fire-detection.


Model: smoke-and-fire-detection-1.0.1.pt
Model size: 127MB
Model type: Object Detection
Framework: PyTorch

Model performance:
Dataset size: 126,293 images
Accuracy: 94.1%
Recall: 93%
F1-Score: 93.5%

Events:
smoke-detected  | Immediate
fire-detected   | Immediate

Event examples:
{
    "scenario": "smoke-and-fire-detection",
    "event_name": "smoke-detected",
    "event_details": {
        "camera": "camera-01",
        "date": "2023-01-04 11:05:02",
        "confidence": 0.92
    }
}
------------------------------------------------

Run a Scenario

Use visionai run command to run a scenario. In its simplest sense, you can run a single scenario on your web-cam. In a more complex use-case, you can specify a pipeline of scenarios, configure notification logic for each scenario, timings for each scenario etc.

$ visionai run --scenario early-smoke-and-fire-detection --camera OFFICE-01

Starting early-smoke-and-fire-detection
...

Next steps

Congratulations! You have successfully run the first scenario. Now go through Tutorials to learn about how to run multiple scnearios, how to configure each scenario for the events you need, how to set up the dependencies etc.

Or you can also go through our scenarios page to explore the different scenarios available and their model details. If you have a need for a scenario to be implemented, do not hesitate to submit a request.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

visionai-0.1.15.tar.gz (74.6 kB view hashes)

Uploaded Source

Built Distribution

visionai-0.1.15-py3-none-any.whl (89.5 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page