As a researcher in the biometric field this is exciting, but has yet to really make any ripples (that i've noticed at least). Face recognition has always been p

As a researcher in the biometric field this is exciting, but has yet to really make any ripples (that i've noticed at least). Face recognition has always been popular; Facebook, Google, etc have their methods and whatnot, the momentum hasn't really changed much there. However, now that biometrics are making their way into more consumer applications I am starting to see more spoofing, liveness tests, and key binding type papers making more of a presence. In my opinion, this is fantastic. A major gripe that i have with the biometric field is that it feels very 'applied' as opposed to 'theoretical'. Let me be clear first, biometrics is an interesting combination of computer vision, machine learning, raw signal/image processing, and matching. Since many pattern recognition/ML researchers rarely read biometric publications (conferences like ICB, IJCB, BTAS, etc. or journals like IEEE TIFS), any good theory that's developed is largely ignored unless the researchers go to places like CVPR or TPAMI (which is very difficult due to a wide array of topics, instead of a smaller more explicit focus). This unfortunately leads to a lack of motivation to focus primarily on theory, instead borrowing what is working in other fields and applying them to our problem. Not a bad thing in any sense of the word, that's what research is - exploring the space of possible solutions. However, it becomes a one way street. We keep looking out, but no one else is looking in. Now back to spoofing, liveness tests, etc. With deep learning, industry has a new spark for ML research. Adding more consumer biometric tech (that aren't developed by Honeywell et al.) means those ML researchers may start looking at journals like TIFS to find engineering suggestions to secure or integrate their systems (e.g. liveness tests). Hopefully a few papers will also catch their interest in the process and maybe...just maybe start to bring more attention to the more focused publications in the field. As the cream rises, we may just see more theoretical results instead stuff of LBP + X method.