Article

Article title FACIAL FEATURES ESTIMATION BY IMAGE MODEL OPTIMIZATION IN SPACE OF THE BASIC FUNCTIONS
Authors A.N. Gneushev
Section SECTION IV. MATHEMATICAL METHODS OF AN ARTIFICIAL INTELLECT
Month, Year 06, 2012 @en
Index UDC 004.93
DOI
Abstract Consideration was given to a problem of the facial features localization in an image. Under investigation [14] the approach of the textural model construction by the learning image set approximation by optimization of the basic function system has been improved. Each facial feature’s texture of the learning set is described by weight vector in the obtained basis and the set of the characteristic point coordinates. The parameters of the object image under consideration are estimated by minimizing the residual of the model weight vectors and projection of the image in the basic space for parameters of the model coordinate system transformation.

Download PDF

Keywords Textural pattern, decomposition optimization; Gabor basic functions; face localization; face tracking; face features extraction; Gabor Wavelet Network.
References 1. Feris R., Krueger V., Cesar Jr.R. Efficient real-time face tracking in wavelet subspace // Proc. IEEE ICCV Workshop on Recog., Anal. and Tracking of Faces and Gestures in Real-time Systems. Vancouver, 2001.
2. Feris R., Gemmell J., Toyama K. et al. Hierarchical wavelet networks for facial feature localization // Proc. IEEE ICCV Workshop on Recog., Anal. and Tracking of Faces and Gestures in Real-time Systems. Vancouver, 2001.
3. Гнеушев А.Н. Построение и оптимизация текстурно-геометрической модели изображения лица в пространстве базисных функций Габора // Изв. АН. Теория и системы управления. – 2007. – № 3. – С. 85–96.
4. Гнеушев А.Н. Оптимизация текстурно-геометрической модели изображения для оценивания параметров лица // Автоматика и телемеханика. – 2012. – № 1. – С. 159-168.
5. Byrd R., Schnabel R.B., Shultz G.A. A trust region algorithm for nonlinearly constrained optimization // SIAM J. Numer. Anal. – 1987. – № 24. – P. 1152-1170.

Comments are closed.