Real-time Human Detection for a Military Company

We love to provide the cutting edge deep learning systems for our customer needs, but our main principle is to make
things that work – quickly and on budget.

Adam Midura

Updated on June 17, 2020 | 2 min read

Detect human contours and center of mass in real-time

One of our customers needed a software solution that detects humans on a video stream which will be as
lightweight as possible to be run on a low cost, small device equipped with an Nvidia GPU.
This made deep learning infeasible for this task. However, we managed to provide a performance-optimized
OpenCV-based algorithm that worked in real-time with the required precision. To speed
up computations, we’ve used the C++ OpenCV library in our Python code – the official Python version
didn’t provide access to CUDA algorithms, so we created a dedicated library to provide appropriate bindings.

Detecting human contours and center of mass in real-time

Further improvements

Currently, we are working on extensions of an existing algorithm to make more comprehensive predictions like
real-time body part segmentation and gesture recognition possible.