Bob’s Skin Color Filter

This module contains the implementation of the skin color filter described in [taylor-spie-2014]. The skin color is modeled as a 2-dimensional gaussian in the normalised rg colorspace. Note that the estimation of the gaussian parameters should be done using a cropped face image. As a consequence, this pacakge depends on the Bob’s face detection package.

Documentation

• User Guide
  • Skin pixels detection in a single image
  • Skin pixels detection in videos
• Python API

References

[taylor-spie-2014]

M.J. Taylor and T. Morris. Adaptive skin segmentation via feature-based face detection, Proc SPIE Photonics Europe, 2014. pdf

Indices and tables

• Index
• Module Index