Machine Vision on FPGA for Recognition of Road Signs

This paper mainly aimed to evaluate the performance of a high level Histogram of Oriented Gradients (i.e., HOG) calculations on labeled image instead of low level block processing on the entire image in pattern recognition and specifically investigate its effectiveness on road sign identification. It was shown that by calculating HOG on the contours of segmented image, not only the final results are promising, but also the amount of processing is significantly reduced and a higher speed is gained.

Autorentext

Ashkan Hashemi was born in Tehran, Iran in 1985. In 2007, he received his Bachelor s degree in Electrical Engineering in Iran. He was awarded the M.Sc degree in Electronics Design from Mid Sweden University, Sundsvall, Sweden in September 2012. Ashkan is currently pursuing his Ph.D degree at Missouri University of Science at Technology, USA.