Printable Version of this PageHome PageRecent ChangesSearchSign In
Ross Beveridge - Face Recognition at CSU and at Large.

I attended Dr. Beveridge's talk and also spoke to him before the colloquium. The fact that i found most interesting about Dr. Beveridge's research is that instead of trying to be at the cutting edge of face recognition technology by developing new face detection algorithms his research focuses on evaluation of these state of the art algorithms. Though personally this work doesnot seem as appealing as actually developing such algorithms, Dr. Beveridge did make a strong case about why such work can be interesting
(it's usefulness for obvious reasons was never in question). He demonstrated various statistical measures developed at CSU to assess the influence of factors such as age,gender and facial expression on algorithm performance. He also spoke about the issue of feature selection for face recognition. The take home message, for me anyway, from the talk was that we are far far away from a general face recognition algorithm and a systematic evaluation of state of the art face recognition technologies can provide important insights towards solving the problem.

Last modified 9 December 2007 at 1:46 pm by Soumya