/CNN_for_SLR

A trained Convolutional Neural Network implemented on ZedBoard Zynq-7000 FPGA.

Primary LanguageVerilogBSD 2-Clause "Simplified" LicenseBSD-2-Clause

CNN_for_SLR

A trained Convolutional Neural Network implemented on ZedBoard Zynq-7000 FPGA. Link to YouTube Video(s): https://www.youtube.com/watch?v=xoB--RFfy6I&feature=youtu.be

Project name: BeeBoard

Date: 30-Jul_2018

Version of uploaded archive: 1

University name: ISTANBUL TECHICAL UNIVERSITY

Supervisor name: Berna Ors Yalcin

Supervisor e-mail: Siddika.ors@itu.edu.tr

Participant(s):

Ilayda Yaman https://www.linkedin.com/in/ilayda-yaman-9bba0ab1/

M. Tarik Tamyurek

Burak M. Gonultas https://www.linkedin.com/in/burak-mert-gonultas-94b045b1/

Email:

ilayda.yaman@eit.lth.se

mttamyurek@gmail.com

gonul004 [at] umn.edu

Board used: Digilent ZedBoard Zynq®-7000 ARM/FPGA SoC Development Board

Vivado Version: 2018.1

Brief description of project: A trained Convolutional Neural Network has been implemented on an FPGA evaluation board, ZedBoard Zynq-7000 FPGA, focused on fingerspelling recognition.

Description of archive (explain directory structure, documents and source files):

CNN folder includes Vivado files

MATLAB_Code folder includes files to verify the results obtained by the Vivado- Behavioral Synthesis

Instructions to build and test project

Step 1: Go to CNN folder for Vivado files of the project

Step 2: Run Behavioral Synthesis

Step 3: Obtain results for the hardware design

Step 4: Compare it with MATLAB results by running the "CNN.m" file inside the MATLAB_Code folder