• <More on Intel.com

Face Recognition: Project

Online semi-supervised learning algorithms incrementally build models of objects, such as faces, and dynamically refine them over time using real-world observations. Researchers on this project team successfully used this approach to develop an effective and highly accurate face recognizer and scaled it to the computational resources of a mobile computer powered by a mainstream Intel® Atom™ processor.

This project complements a number of other Intel Lab research projects. The identity data can be used as an input for other context-aware computing applications, providing a reliable mechanism for personalizing apps and services for individuals. This project has already spawned practical applications, and the C++ code for the recognition and self-learning modules has been handed off to other groups at Intel for further development. Online semi-supervised learning also has broader applications that address the capabilities of a computing device to adapt itself to the behaviors of a user.

Research Agenda

  • Improve the quality of real-time machine learning on ultra mobile computing devices without asking for feedback (interaction with the user of the device)
  • Develop algorithms that can progressively build a model of objects and enhance it over time
  • Improve the accuracy of algorithms that perform face recognition

Recent Publications

Michal Valko, Branislav Kveton, Ling Huang, and Daniel Ting. Online Semi-Supervised Learning on Quantized Graphs. In Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence, Catalina Island, California, July 2010.

Branislav Kveton, Michal Valko, Matthai Philipose, and Ling Huang. Online Semi-Supervised Perception: Real-Time Learning without Explicit Feedback. In Proceedings of the 4th IEEE Online Learning for Computer Vision Workshop, San Francisco, California, June 2010. Best paper award.

Branislav Kveton, Michal Valko, Ali Rahimi, and Ling Huang. Semi-Supervised Learning with Max-Margin Graph Cuts. In Proceedings of the 13th International Conference on Artificial Intelligence and Statistics, Sardinia, Italy, May 2010.

Learn More

Video showing face-tracking and enhanced speech recognition techniques >

Video highlighting an application that demonstrates face recognition using Intel® Media SDK sample code >

Video offering an example of low-power mobile computer possibilities for face recognition >