Beauty and Wellness

Beauty and wellness are areas where one usually would not expect applications from digital image processing and computer vision. However, our research shows that these fields offer an interesting, so far unexplored potential.

Image registration is the task of computing the deformation field that transforms a first image into a second one that depicts the same content. Typically, registration is used in the context of medical imaging, where one is interested in matching image data obtained by different modalities such as computer tomography (CT) or magnetic resonance tomography (MRT). If used for hairstyle simulation, however, registration serves a completely different purpose: One seeks the deformation field that allows to map a reference hairstyle onto a customer face. Thus it becomes possible to select the perfect hairstyle right from the start. The rest is up to the hairdresser ...

Customer Face
Reference Face / Hairstyle
Deformation Field
Final Result

  • Hairstyle Simulation with Variational Methods.
    One of the few applications in our group mainly dedicated to women is the simulation of hairstyles. Using our expertise in correspondence problems, we have developed an algorithm that allows to automatically register a chosen hairstyle onto a customer face [1]. By preregistering the eyes and masking out the haircut, we are able to compute the deformation field that is required to perfectly fit a given reference hairstyle to a photographed customer face. This algorithm is actually used in practice: It is part of HairConcept - a software from Style Concept that is very popular among hairdressers.

  1. O. Demetz, J. Weickert, A. Bruhn, M. Welk:
    Beauty with variational methods: An optic flow approach to hairstyle simulation.
    In F. Sgallari, A. Murli, N. Paragios (Eds.), Scale Space and Variational Methods in Computer Vision. Lecture Notes in Computer Science, Vol. 4485, 825-836, Springer, Berlin, 2007.

MIA Group
The author is not
responsible for
the content of
external pages.

Imprint - Data protection