Driver drowsiness detection using back-propagation neural network / Endratno Ibrahim

Faces as the primary part of human communication have been a research target in a
computer vision over a few decades. This project focuses on the development of Back
propagation neural network for driver drowsiness detection based on eyes state (open
and close). It uses a CCD camera equipped with an active IR illuminator to acquire
images of the driver. Then the images sequence will be process offline to determine
the drowsiness. This project will provides the confirmation that back propagation is
suitable for this type of system. There are two important phases that were focused in
this system development. The phases are pre-processing phases and neural network
design phase. Every phase has a several sub processes and the network parameter are
the predetermine values in the training process. Several suggestions and
recommendations are proposed to enhance the detection presence and performance.