Files in this item



application/pdfECE499-Sp2014-yuYangyang.pdf (1MB)Restricted to U of Illinois
(no description provided)PDF


Title:An Implementation of a Human Position Recognition System on FPGA
Author(s):Yu, Yangyang
Contributor(s):Chen, Deming
Subject(s):human position recognition system
dynamic vision sensors
Abstract:The thesis describes an implementation of a human position recognition system on FPGA (Filed Programmable Gate Array) using event-based dynamic vision sensors. Different from conventional vision sensors that detect full frames of images, the dynamic vision sensors catch pixels that changed intensity, which is more similar to the way human retina works. By combining the new sensor technology and the FPGA technology, the real-time configurable human position recognition system is able to achieve both higher performance and better energy efficiency. A bio-inspired cluster detection algorithm is used for figure detection, a line segment extraction algorithm is used for feature extraction and a Hausdorff distance algorithm is used for classification. This paper presents the system design and the FPGA implementation of the input interface and the cluster detection unit.
Issue Date:2014-05
Date Available in IDEALS:2014-10-31

This item appears in the following Collection(s)

Item Statistics