How can you deploy the PPE detection solution?
It's easier than it may seem. We have prepared the answers for the frequently asked questions about this solution.
You seem to have a solution for my company. How do I get started with it?
Welcome, dear guest! Our team aims to offer responsive, competent and excellent service. So first of all let us invite you to an intro-call, where we can dive deeper into your business objectives and figure out the way we can provide you the best solution. Feel free to ask any questions, we’ll do our best to help you.To request a free, however full-fledged consultation, proceed here.
What steps will be taken from now till the full deployment? How long will it take?


What deployment options are available?
PPE Monitoring Platform is available by subscription in the following options:


I might need customization or support. What about this?
We understand that turbulent markets may establish new priorities for business as the time goes by. Moreover special functions of a product can expand business opportunities and hence provide a significant competitive advantage. Therefore our team is ready to serve your company and provide consistent solutions that fit your goals.
Our Advanced Support Services include:
-
system monitoring
-
collection of the additional dataset
-
model retraining for enhancements once every 3 months
-
hot and priority fixing bugs and system problems
-
advanced training.
If any additional options are needed, it's all up for discussion. Contact us anytime and for any questions.
Some examples of possible customization:
-
Adding your custom PPE items to be monitored
-
Integration of the solution with with you systems (access control, social distancing, etc)
-
Adding some new functions and capabilities (color detection, size detection, etc)
Or you can request an individual customization, and we will do our best to meet your needs.

How does the solution work on AWS Marketplace?
The PPE Monitoring Platform consists of two components:
The ML model itself, which analyzes images from live stream from client’s surveillance cameras on the worksites;
The Web interface, which transfers images from cameras to the ML model for analysis. The interface is user-friendly, so it serves as the main workplace environment for an end user. It provides an ability to:
-
control the cameras;
-
observe the process of video stream analysis in real-time;
-
get and configure notifications about detected non-compliance cases:
-
get access to all the detected violations, automatically saved to an archive;
manage users and accesses.
The architecture on AWS Marketplace is depicted below:


Have any questions? We are here to help.
What are the technical requirements for an ML model server and web-interface server?

For ML server:
-
Server with Docker and NVIDIA drivers installed
-
GPU: NVIDIA K80/P100/V100/A100/T4 or similar
-
CPU: Intel Xeon min 2.0 GHz
-
RAM: min 16 GB
-
Storage: min 100GB
-
EDGE platforms like NVIDIA EGX A100 or NVIDIA Jetson
For web-interface server:
-
Storage: min 1TB (SSD)
-
CPU: Intel Xeon min 1.3 GHz
-
RAM: min 8 GB
What are requirements for cameras and monitored objects?
-
Please note, that VITech Lab does not provide the needed hardware, we use your existing cameras. If this is a kind of a barrier, we have partners who can provide you with cameras.
-
Resolution - from 2 MP
-
Video resolution - from 1920 × 1080 pixels
-
Video Format - H.264
-
Max FOV: 80 degrees
-
Color
-
The size of the analyzed object - a minimum 5% of the camera image size

What are requirements for cameras and monitored objects?
-
The ML model speed is up to 20 fps (V100-based server). It also depends on the cameras’ image resolution and the number of people on each image.
-
Another important thing about speed is that 1 server can work with 10 cameras in real-time, while one stream processing runs at a speed of 10 fps (V100-based server). The number of processed threads can be increased by reducing the processing speed.

How do you guarantee security for my data?
If you choose to deploy the solutions in cloud, using your AWS Marketplace account, then be sure that AWS Marketplace itself meets high security level and has the appropriate certificates. In such case we do not have access to your data. For more details visit:
If you choose an option where the solution is deployed locally, then all your data is kept locally, and we also do not have the access to it.