next up previous
Next: ch4 Up: ch4 Previous: ch4

Neural Nets for Face Recognition




\begin{picture}(200,200) %% a pt = 1/72 in
\put(35,20){\psfig{file=./bookps/mitc...
...(134,110){\mbox{\small 30x32}}
\put(134,100){\mbox{\small inputs}}
\end{picture}

\fbox{\psfig{file=./bookps/glickman-left-neutral-open-4.ps,width=1.3 in}} \fbox{\psfig{file=./bookps/mitchell-straight-happy-sunglasses-4.ps,width=1.3 in}} \fbox{\psfig{file=./bookps/bpm-right-angry-open-4.ps,width=1.3 in}} \fbox{\psfig{file=./bookps/saavik-up-neutral-sunglasses-4.ps,width=1.3 in}}
Typical input images





90% accurate learning head pose, and recognizing 1-of-20 faces



Don Patterson 2001-12-13