Integration of precise iris localization into active appearance models for automatic initialization and robust deformable face tracking

Face tracking and facial landmarking plays an important role in human-machine interaction. Active-Appearance Models (AAM) represent a well-established method to mathematically describe face characteristics such as facial expression and attentiveness. However, the applicability of AAM suffers from their high sensitivity to initialization and low robustness in scenarios of strong head movements. We show how a robust, automatic initialization of AAM can be performed by integration of precise iris localization into the AAM fitting algorithm. The shape and global transform parameters are extracted from the re-initialized model to provide well-suited start parameters for the subsequent fitting. The method is validated on the gi4e database showing an improvement to naive initialization and re-initialization in terms of robustness and accuracy.

[1]  Xiaoming Liu,et al.  Generic Face Alignment using Boosted Appearance Model , 2007, 2007 IEEE Conference on Computer Vision and Pattern Recognition.

[2]  Fernando Puente León,et al.  Combining isophote and cascade classifier information for precise pupil localization , 2016, 2016 IEEE International Conference on Image Processing (ICIP).

[3]  Takeo Kanade,et al.  Real-time combined 2D+3D active appearance models , 2004, CVPR 2004.

[4]  Maja Pantic,et al.  Active Orientation Models for Face Alignment In-the-Wild , 2014, IEEE Transactions on Information Forensics and Security.

[5]  Qiang Ji,et al.  Shape Augmented Regression Method for Face Alignment , 2015, 2015 IEEE International Conference on Computer Vision Workshop (ICCVW).

[6]  Rafael Cabeza,et al.  A novel 2D/3D database with automatic face annotation for head tracking and pose estimation , 2016, Comput. Vis. Image Underst..

[7]  Qingshan Liu,et al.  Facial Shape Tracking via Spatio-Temporal Cascade Shape Regression , 2015, 2015 IEEE International Conference on Computer Vision Workshop (ICCVW).

[8]  Stefanos Zafeiriou,et al.  A Comprehensive Performance Evaluation of Deformable Face Tracking “In-the-Wild” , 2016, International Journal of Computer Vision.

[9]  David Cristinacce,et al.  Automatic feature localisation with constrained local models , 2008, Pattern Recognit..

[10]  Timothy F. Cootes,et al.  Coupled-View Active Appearance Models , 2000, BMVC.

[11]  Timothy F. Cootes,et al.  Multi-view Constrained Local Models for Large Head Angle Facial Tracking , 2015, 2015 IEEE International Conference on Computer Vision Workshop (ICCVW).

[12]  Ralph Gross,et al.  Generic vs. person specific active appearance models , 2005, Image Vis. Comput..

[13]  David J. Kriegman,et al.  Localizing parts of faces using a consensus of exemplars , 2011, CVPR.

[14]  Stefanos Zafeiriou,et al.  HOG active appearance models , 2014, 2014 IEEE International Conference on Image Processing (ICIP).

[15]  Maja Pantic,et al.  Fast and exact bi-directional fitting of active appearance models , 2015, 2015 IEEE International Conference on Image Processing (ICIP).

[16]  Stefanos Zafeiriou,et al.  The First Facial Landmark Tracking in-the-Wild Challenge: Benchmark and Results , 2015, 2015 IEEE International Conference on Computer Vision Workshop (ICCVW).

[17]  Timothy F. Cootes,et al.  Active Appearance Models , 2001, IEEE Trans. Pattern Anal. Mach. Intell..

[18]  Simon Baker,et al.  Active Appearance Models Revisited , 2004, International Journal of Computer Vision.

[19]  Ashraf A. Kassim,et al.  Facial Landmark Detection via Progressive Initialization , 2015, 2015 IEEE International Conference on Computer Vision Workshop (ICCVW).

[20]  Maja Pantic,et al.  Gauss-Newton Deformable Part Models for Face Alignment In-the-Wild , 2014, 2014 IEEE Conference on Computer Vision and Pattern Recognition.