Back to list

Face Reconstruction


Model for restoring blurred and pixelated faces in a photo.

The model is built on a base of a trainable neural network that receives an image with a blurred or pixelated face, recognizes and analyzes it, and then restores it as close to the original state as possible.

Our developers have processed a training images dataset and analyzed the results within several cycles. Eventually we have a full-value API that can restore the face in a photo highly realistically.

At the moment, the model is already showing decent results.

Desktop - 4.png
Development time
8 weeks 3 developers