The project is about some fundamental pattern/face recognition problems. One example (and there are many more) is theoretical  analysis of kernel functions/models, such as fractional power polynomial models.  Analyze and predict the behavior of kernel functions/models, and develop new models with better pattern recognition performance.

The project requires good math in probability and statistics (knowledge of multivariate statistical analysis and functional analysis is a plus).
Good publication record is preferable, as it is expected that several good journal submissions should be made each year.

Salary will be $35,000 per year, and the starting date is October. Interested persons please send CV and contacts of references to  chengjun.liu@njit.edu.