The naked truth (estimating body shape under clothing)

Remember ‘X-Ray Specs’? “Look at your friend. Is that really his body you ‘see’ under his clothes?” asked the advertisements. Though no doubt disappointing to some, the answer to the question was of course ‘No’. Since their invention however, progress has been made in computerised body visualisation systems, devices which also can’t ‘see’ under peoples clothes, and yet, in a curious kind of way (as the video above shows), can.

For background, there’s probably no better place to start than the website of professor Michael Black of the Max Planck Institute for Intelligent Systems (Perceiving Systems), Tübingen. Germany. The professor began publishing research details in 2008 (along with Dr. Alexandru O. Bălan ) see: ‘The naked truth: Estimating body shape under clothing’ In European Conf. on Computer Vision, ECCV, Springer-Verlag, volume 5304, LNCS, pages 15-29, Marseilles, France. October 2008.

“We propose a method to estimate the detailed 3D shape of a person from images of that person wearing clothing. The approach exploits a model of human body shapes that is learned from a database of over 2000 range scans. We show that the parameters of this shape model can be recovered independently of body pose. We further propose a generalization of the visual hull to account for the fact that observed silhouettes of clothed people do not provide a tight bound on the true 3D shape. With clothed subjects, different poses provide different constraints on the possible underlying 3D body shape. We consequently combine constraints across pose to more accurately estimate 3D body shape in the presence of occluding clothing. Finally we use the recovered 3D shape to estimate the gender of subjects and then employ gender-specific body models to refine our shape estimates.“

Applications :

“There are many applications and future directions for this new line of research. For human tracking in video it may be useful to estimate limb lengths, body shape parameters and body mass as these could be used in the inference of dynamics. Body shape parameters could be used in visual tracking applications to identify and re-acquire subjects who come in and out of the field view. For forensic video applications, the extraction of body shape parameters could be useful in identifying suspects. There are also many applications of these methods in personal fitness, retail apparel and computer games.”

Privacy concerns :

“Privacy concerns must be addressed for any technology that purports to ‘see’ what someone looks like under their clothes. Unlike backscatter X-ray and infra-red sensors, our approach does not see through clothing. It does not have any information about the person’s body that is not available essentially to the naked eye; in this sense it is not intrusive. [orig. emphasis]

Note: The work was supported in part by the Office of Naval Research.

Update: Since 2008 many other research teams across the globe have been looking into the possibilities of algorithmic clothing removal. For a recent example, see W.-Y. Chang and Y.-C. F. Wang, ‘Seeing Through The Appearance: Body Shape Estimation Using Multi-View Clothing Images’, IEEE International Conference on Multimedia & Expo (ICME), June 2015.

Bonus: If you have WebGl™ running on your internet browsing system (it’s a built-in component of many modern browsers, but you may have to enable it) then you can view an interact in real-time with the BodyVisualiser provided by Max Planck Gesellschaft.