Faces

Iterait's AI opens the world of marketing to new opportunities. Faces AI allows in the real-time target offline marketing campaigns to the right audience and measure the performance of your campaigns in the same time. We are happy to deliver this project to our business partner AD 24.

Learn more

Contribution

Faces is an artificial intelligence for tracking people in a video and predicting their age and gender.

Compared to other similar applications, it does not need a dedicated server with expensive GPU cards - it needs only Raspberry Pi.

However, even on this cheap hardware, it achieves competitive performance - the F1 score for detecting people on widely used Wider benchmark is 0.7 which corresponds to the ability to detect people up to 9 meters far away from the camera.

The accuracy of gender prediction is 95%; the average error of age prediction is less than seven years.

Since it was trained on our specifically annotated dataset, it works even in suboptimal conditions such as bad light or fast moving people.

Technology

Faces project utilizes the latest advancements in deep neural networks and computer vision algorithms. On top of that, we have updated specially for this project our Emloop framework that now enables to train powerful AIs for edge computing devices like Raspberry Pi.