A solution using neural networks allows you to identify certain objects in images or video. For example, it can be used to analyze images of people to target social media ads.

The detectors operate at a speed of about 100 frames per second on a 640x480 video stream. Accuracy and completeness depend on the task at hand and the required speed. A modification of the deep architecture of ResNet, the PuzzleLib neural network framework was used. The detector works on NVIDIA GTX 1080 graphics cards.

The solution has an API and is also provided as an online service.