The overall aim of this problem is to use two digital photographs of a person to compute the person’s 3-D shape. The photographs are taken with ‘ordinary’ equipment; a monocular lens in a smartphone, or a web-camera in a lap-top computer would be typical. There is no specification on the brand of equipment.
The individual is requested to wear close-fitting clothing and to adopt particular poses for the two photographs, at a distance of about 3m from the camera. One image is face-on to the camera, and the other is side-on to the camera. Poikos ask the person his/her height for a sense of scale.
Poikos have broken the problem of deriving the 3-d shape from the 2-d photographs into a series of smaller problems. The first of which is to find the edge of the person in an image. That is, they wish to separate the person, “foreground”, from the rest of the image, “background”. This is achieved by comparing the image against a pre-calculated probability map for the pose, combined with colour information from the image and edge detection. Once the 2D silhouette for each input image is identified, it is compared against known 3D shapes using a similar system to 'nearest match' to convert the silhouettes into a three dimensional scan. Once a good fit is found, refinements are made to tune the shape to the subject.
The challenge for the Study Group is to investigate methods of combining the image segmentation and 3D reconstruction stages. Furthermore, Poikos are interested in using the images alone to estimate parameters relating to the scale of the image and several types of camera defect.
Problem presented by:
David Evans (Poikos)
Eleanor Watson (Poikos)