Purdue University Graduate School
Browse

Region-based Convolutional Neural Network and Implementation of the Network Through Zedboard Zynq

Download (5.15 MB)
thesis
posted on 2019-06-10, 16:31 authored by MD MAHMUDUL ISLAMMD MAHMUDUL ISLAM
In autonomous driving, medical diagnosis, unmanned vehicles and many other new technologies, the neural network and computer vision has become extremely popular and influential. In particular, for classifying objects, convolutional neural networks (CNN) is very efficient and accurate. One version is the Region-based CNN (RCNN). This is our selected network design for a new implementation in an FPGA.

This network identi es stop signs in an image. We successfully designed and trained an RCNN network in MATLAB and implemented it in the hardware to use in an embedded real-world application. The hardware implementation has been achieved with maximum FPGA utilization of 220 18k_BRAMS, 92 DSP48Es, 8156 FFS, 11010 LUTs with an on-chip power consumption of 2.235 Watts. The execution speed in FPGA is 0.31 ms vs. the MATLAB execution of 153 ms (on computer) and 46 ms (on GPU).

History

Degree Type

  • Master of Science in Electrical and Computer Engineering

Department

  • Electrical and Computer Engineering

Campus location

  • Indianapolis

Advisor/Supervisor/Committee Chair

Dr. Lauren Christopher

Additional Committee Member 2

Dr. Maher Rizkalla

Additional Committee Member 3

Dr. Paul Salama