3D "Before-And-After" Tool for Plastic Surgeons

Innovative software based on real clinical data helps plastic surgeons show their patients a three-dimensional (3D) anatomically accurate after-surgery image.

Researchers at Tel Aviv University (TAU, Israel) and Stanford University (CA, USA) developed "Shape Google”, a complex mathematical computer modeling tool that predicts "deformations” of nonrigid objects, by retrieving geometric objects in the same manner that the Google search engine retrieves web pages. This approach allows the representation of the images as collections of "visual words.” By constructing compact and informative shape descriptors using the "bag of features” approach, the researchers showed that considering pairs of "geometric words” allows creating spatially sensitive bags of features with better discriminative power. Adopting metric learning approaches, the shapes could then be efficiently represented as binary codes.

The software tool helps the patients avoid unexpected results in the plastic surgeon's office, and can also help a surgeon determine the most favorable outcomes. Following rigorous interviews with internationally respected plastic surgeons, the researchers designed the program with the help of numerous pre- and post-surgery images fed into a computer to train it to generate post-surgery images more accurately, by integrating multiple 2D images into a 3D output. The study describing the new approach was published in the January 2011 issue of the Journal of the Association for Computing Machinery (ACM) Transactions on Graphics.

"Our program is more like a virtual mirror. It gives surgeons and their patients a way to see a 3D before-and-after image as though the patient has really undergone the operation,” said lead developer Alex Bronstein, PhD, of the TAU department of electrical engineering. "The same premise can be used by people in weight-loss programs - as a predictor of their body image after they've shed excess pounds.”

Related Links:
Tel Aviv University
Stanford University