Autors: Martin Penev., Manolova, A. H., Boumbarov, O. L.
Title: Active Shape Models with 2D profiles for Stress/Anxiety recognition from face images
Keywords: Active Shape Model, facial emotion recognition

Abstract: The facial expression is a visible manifestation of the emotional state, cognitive activity, intention, personality, and mental health of a person. In this paper we use an approach to distinguish between different emotions from images in particular the unhappiness related to anxiety and stress. The approach is based on Active Shape Model (ASM) with 2D profile for extracting important facial points (coordinates of edges of the mouth, eyes, eyebrows, etc.). Further the feature vectors are formed by concatenating the landmarks data from the proposed method and use the data as an input to kernel Support Vector Machine (SVM) classifier. The experimental results using Cohn-Kanade Extended Facial Expression Database show high recognition rate. 2D profile ASM significantly speeds up the fitting process comparing with 1D profile ASM by averagely over 40%.

References

  1. [2] G. Donato, M.S. Bartlett, J.C. Hager, P. Ekman, T.J. Sejnowski, 1999, Classifying Facial Actions, Pattern Analysis and Machine Intelligence, Volume 21(10), pp. pp. 974-989
  2. [10] A. Lanitis, C.J. Taylor, and T.F. Cootes, 1997, Automatic Interpretation and Coding of Face Images Using Flexible Models, Pattern Analysis and Machine Intelligence, Volume 19(7), pp. pp. 743-756

Issue

9th International Conference on Communications, Electromagnetics and Medical Applications, vol. 1, pp. 108-112, 2014, Bulgaria, ISSN 1314-2100

Full text of the publication

Вид: пленарен доклад в международен форум