PHOTO-editing software that knows what your friends and family look like could transform blurry snaps of them into pictures worth keeping, according to computer vision engineer at the University of California, San Diego.
When Joshi and his colleagues studied home photo collections they saw that many pictures of faces were blurry. 鈥淧recious moments were often lost due to blur induced by camera shake and poor lighting,鈥� he says.
Advertisement
The team then wrote software to solve the problem (). Their algorithm uses facial recognition methods like those in some photo editors to find a sharp image with a similar pose. The pattern of colour and light in the blurry face is then tweaked to match the model photo.
