Integral Projections based face tracking algorithm provides high accuracy and computational efficiency to accurately track human faces. However, the inherent complexity of having three dimensions and the usage of projection model with each dimension has posed a challenge to deploy on resource constrained handheld devices. Motivated by these observations, we propose a multi-threaded architecture on Android to address this issue. First, we discuss few related works to investigate and analyze the Integral Projection face tracking algorithm and detail the constraints associated to identify face tracking accurately. Second, we discuss few characteristics of android architecture and detail constraints associated with middleware sub-system to cater face tracking efficiently. Consequently, we describe the proposed architecture and the rationale of functionality based module classification to address inherent challenges like reducing false positive rate on resource constrained devices. In particular, we point out the benefits of the Face Algorithm Agent to ensure smooth and swift portability onto different flavors of android. To validate our approach, trial experiments were performed using the said architecture on Zoom2 hardware platform based on Android. Experimental results lead us to conclude the architecture is efficient at reducing the false positive rates significantly. Finally, future research directions based on the current architecture results are pointed out.

1. Refining Face Tracking with Integral Projections Gin´es Garc´ıa Mateos Dept. de Inform´atica y Sistemas, Universidad de Murcia 30.170 Espinardo, Murcia (Spain).
2. Bradski, G. R.: Computer Vision Face Tracking For Use in a Perceptual User Interface. Intel Technology Journal Q2’98 (1998)
3. Spors, S., Rabenstein, R.: A Real-Time Face Tracker for Color Video. IEEE Intl. Conference on Acoustics, Speech, and Signal Processing, Utah, USA (2001)
4. Kaucic, R., Blake, A.: Accurate, Real-Time, Unadorned Lip Tracking. Proc. of 6th Intl. Conference on Computer Vision (1998) 370–375
5. Sobottka, K., Pitas, I.: Segmentation and Tracking of Faces in Color Images. Proc. of 2nd Intl. Conf. on Aut. Face and Gesture Recognition (1996) 236–241
6. Stiefelhagen, R., Yang, J.,Waibel, A.: A Model-Based Gaze Tracking System. Proc. of IEEE Intl. Symposia on Intelligence and Systems (1996) 304–310
7. Pahor, V., Carrato, S.: A Fuzzy Approach to Mouth Corner Detection. Proc. Of ICIP-99, Kobe, Japan (1999) I-667–I-671
8. Schwerdt, K., Crowley, J.L.: Robust Face Tracking Using Color. Proc. of 4th Intl. Conf. on Aut. Face and Gesture Recognition, Grenoble, France (2000) 90–95
9. Garc´ıa-Mateos, G., Ruiz, A., L´opez-de-Teruel, P.E.: Face Detection Using Integral Projection Models. Proc. of IAPR Intl. Workshops S+SSPR’2002, Windsor,Canada (2002) 644–653
10. Isard, M., Blake, A.: Contour Tracking by Stochastic Propagation of Conditional Density. Proc. 4th Eur. Conf. on Computer Vision, Cambridge, UK (1996) 343–356
11. Vieren, C., Cabestaing, F., Postaire, J.: Catching Moving Objects with Snakes for Motion Tracking. Pattern Recognition Letters, 16 (1995) 679–685
12. Pentland, A., Moghaddam, B., Starner, T.: View-Based and Modular Eigenspaces for Face Recognition. Proc. CVPR’94, Seattle, Washington, USA (1994) 84–91
13. La Cascia, M., Sclaroff, S., Athitsos, V.: Fast, Reliable Head Tracking Under Varying Illumination: An Approach Based on Registration of Texture-mapped 3D Models. IEEE PAMI, 22(4), (2000) 322–336
14. Intel Corporation. IPL and OpenCV: Intel Open Source Computer Vision Library. http://www.intel.com/research/ mrl/research/opencv/
15. Kaijian Xu, Manli Zhu, Daqing Zhang, Tao Gu: “Context-Aware Content Filtering & Presentation for Pervasive & Mobile Information Systems“; Proceedings of the 1st international conference on Ambient media and systems, Quebec, Canada, 2008.

Face Tracking Android Architecture QoE