Researchers from the Chinese Academy of Sciences in Beijing have used artificial intelligence to develop software that they call ‘DeepFaceDrawing’ that is capable of producing photo-realistic images of people with nothing more than a rough sketch.
As reported by Engadget, instead of having to rely on detailed examples of eyelids or lip shapes, the software uses a combination of probability and a database compiling of faces and facial components. It then considers how the different facial elements like eyes, nose, mouth, face shape, and hair type work with each other to construct an image.
According to the research paper documenting the progress made in this experiment, “Recent deep image-to-image translation techniques allow fast generation of face images from freehand sketches. However, existing solutions tend to overfit to sketches, thus requiring professional sketches or even edge maps as input.
“To address this issue, our key idea is to implicitly model the shape space of plausible face images and synthesize a face image in this space to approximate an input sketch. Our method essentially uses input sketches as soft constraints and is thus able to produce high-quality face images even from rough and/or incomplete sketches.”
The code for DeepFakeDrawing is “coming soon” according to the project’s website, so it might not be too long before we potentially see it being used by law enforcement to quickly generate images of suspects based on nothing more than a sketch.
Take a look at the DeepFaceDrawing in action below:
Image: Chinese Academy of Sciences