NCTU_DefectDetection

arc_design_contest/2019/NCTU_DefectDetection/README.md

@@ -0,0 +1,68 @@
+# Eletronic Component Defect Detection on ARC IOT Development Kit
+* [Introduction](#introduction)
+    * [System Architecture](#system-architecture)
+* [Hardware and Software Setup](#hardware-and-software-setup)
+    * [Required Hardware](#required-hardware)
+	* [Required Software](#required-software)
+	* [Hardware Connection](#hardware-connection)
+* [User Manual](#user-manual)
+	* [Before Running This Application](#before-running-this-application)
+	* [Run This Application](#run-this-application)
+## Introduction
+In this contest, we aim to develop a convolutional neural network model to detect the defect on eletronic component, such as capacitor and resistor. To run on the ARC Development Kit (IOTDK), we need to create a model with small amount of paramter.
+
+### System Architecture
+![](https://imgur.com/VJ883mq.png=650x)
+
+We get a dataset with three kinds of data, defect, false-alarm, and golden. False-alarm and golden are kinds of good product, but defect is a kind of defective product. Then, we have a 2-class classifier convolutional neural network which is trained with the training images. Finally, we get a result which is good or defect.
+
+## Hardware and Software Setup
+### Required Hardware
+[DesignWare ARC IoT Development Kit](https://github.com/foss-for-synopsys-dwc-arc-processors/ARC-Development-Systems-Forum/wiki/ARC-Development-Systems-Forum-Wiki-Home#arc-iot-dev-kit-1)
+![](https://imgur.com/kGNpoki.png=500x)
+
+| Training Platform     | Device                                               |
+| --------------------- | ---------------------------------------------------- |
+| CPU                   | Intel(R) Xeon(R) CPU E5-2620 v4 @ 2.10GHz (32 Cores) |
+| GPU                   | GeForce GTX 1080                                     |
+| RAM                   | 188GB                                                |
+| OS                    | Linux Ubuntu 14.04.5                                 |
+
+### Required Software
+* EMSK Environment 
+	* [embARC OSP branch-embARC MLI](https://github.com/foss-for-synopsys-dwc-arc-processors/embarc_osp/tree/embarc_mli)
+	* [Teraterm](https://ttssh2.osdn.jp/index.html.en)
+	* [Environment Build Up](https://embarc.org/embarc_osp/doc/build/html/getting_started/getting_started.html)
+* Python 3.7
+* Deep Learning related Package
+    * Jupyter (optional)
+    * Caffe
+	* Numpy
+	* Skimage
+	* Matplot
+
+### Hardware Connection
+* IOTDK connection
+
+![](https://imgur.com/Z9IYyTH.png=400x)
+
+## User Manual
+### Before Running This Application
+* Dataset
+
+Due to ownership, we could not provide dataset. You should prepare it as a char array by yourselves.
+
+* makefile
+```
+BOARD ?= iotdk
+CUR_CORE ?= arcem9d
+APPL_DEFINES = -DUSE_APPL_MEM_CONFIG -DV2DSP_XY -DMODEL_BIT_DEPTH=8
+EMBARC_ROOT = setting your root path of your embARC-osp without " "
+```
+
+### Run This Application
+Run the "make run" command on cmd in directory where your makefile is
+![](https://imgur.com/ybhg4RR.png)
+
+Result:
+![](https://imgur.com/cL7dEJy.png)