Active Shape Models with Stasm

after-stasm-small.jpg before-stasm-small.jpg

Stasm is a C++ software library for finding features in faces. You give it an image of a face and it returns the positions of the facial features.

Stasm is designed to work on front views of faces with neutral expressions. It performed very well in an independent 2013 comparative study.

Source code is provided under a BSD style license. OpenCV is required.

Source code example.

Wasm utility to try out Stasm on a Windows system. (This can be easily uninstalled by running the included uninstaller, or from "Control Panel | Add or Remove Programs", leaving no trace on your system.)

stasm4vs3.png

Download

There are basically two variants of Stasm. The latest version of these variants is given below.

Documentation

Stasm is based on the Active Shape Model of Tim Cootes and his colleagues. The algorithms are described in the following references:

Citing Stasm

Please cite the following paper in any publicly available text that uses Stasm:
@article{Milborrow2014,
  author={S. Milborrow and F. Nicolls},
  title={{Active Shape Models with SIFT Descriptors and MARS}},
  journal={VISAPP},
  year={2014}
}
Note that our 2008 ECCV paper no longer applies to this version of Stasm.

Additional experiments

How reliable are manual landmarks? To get an idea of the reliability of manual landmarks, we did a small experiment.

Can we estimate the quality of the landmark positions generated by the Active State Model? Alas, the short answer is not easily. See this experiment.

Acknowledgments

I would like to thank the following people who provided ideas, code, data, and techniques used in Stasm.
Back to Stephen Milborrow Homepage