Lecturers‎ > ‎

Terence Sim

Title:  Face Recognition

Speaker:  Dr. Terence Sim, Assistant Professor, School of Computing, National University of Singapore, Singapore


Humans are very good at face recognition.  We therefore expect algorithms to do just as well, but this is still quite far off.  To be sure, there are many commercial face recognition systems that perform quite well at selected tasks, but overall, the problem remains challenging and exciting.  In these two lectures, we will examine the difficulties in automatic face recognition in its different applications.  We will also highlight 5 key scientific papers spanning 20 years that have made significant contribution to the field.  

Next, we turn out attention to human performance: exactly how well do humans recognize faces?  Understanding these quirks will give us insight into our mental processes, and may yet lead to better algorithms.  We will then conclude with a comparison of how algorithms and humans perform.

Topics to be covered:

* Appearance variations
* Applications
* 5 key papers
* Human Performance