A group of faculty members from the University of Washington has developed a new algorithm for transforming 2D static images into 3D moving models. The starting material is photographs, which the “Photo Wake-Up” technology makes “come to life”, in front of the viewer's eyes, leave the frame and start performing different movements in the surrounding world. The development will surely be a breakthrough in the field of augmented reality.
For the algorithm to work, it needs a picture on which the animation object can be more or less clearly distinguished. It is not so important whether it is a real photograph, children's scribbles, a frame from an anime, or even a Picasso painting, the main thing is the presence of a humanoid object. The algorithm recognizes body parts, separates them from the general composition, models a 3D skeleton based on proportions and mounts a 2D texture for it, copying the original drawing.
Further, with such a model, you can perform arbitrary procedures, for example, give the command to run to another picture in the gallery and swap places with the character from there. It looks like magic or, more precisely, as a vivid and visual application of augmented reality technologies. Unlike its predecessors, the authors of this algorithm managed to achieve unprecedented accuracy in recognizing and copying parts of the original picture. And so the viewer easily believes in an illusion.