Both these concerns are addressed by porting existing
automotive vision algorithms onto a GPU. Section II discusses
the algorithms that are evaluated on the GPU. The Viola and
Jones object detection algorithm is mainly used for pedestrian
and vehicle detection applications and the Morphology-based
region extraction algorithm is used to help extract distinct
shapes. In Section III we will discuss the necessity to use the
data obtained from the above methods and apply it to
performance characterization. Section IV will talk about the
data flow of the system. Section V will talk about the
acceleration possible on GPU architecture. In Section VI we
will conclude with possible future work on this subject.