Project information

  • Category: Android Apps
  • Project date: Nov 2022

RND Apps


Tools :

Project Story

This study aims to implement the Convolutional Neural Network into an android-based application that can detect and recognize symbols of electrical components in a short time. This application uses the Convolutional Neural Network technique which is designed to process a collection of digital image data. This app is motivated by the difficulty of electrical students in recognizing or knowing the symbols of electrical components, especially symbols that have similar shapes.

Data Set Table


The Process


Result

  • Normal Conditions:
    • Printed on Paper Media: 100% accuracy
    • Handmade Symbols: 98.5% accuracy
  • Rotation of 10 Degrees:
    • Printed on Paper Media: 92.6% accuracy
    • Handmade Symbols: 85.3% accuracy
  • Rotation of -10 Degrees:
    • Printed on Paper Media: 95.6% accuracy
    • Handmade Symbols: 100% accuracy
  • Horizontal Flipping:
    • Printed on Paper Media: 27.9% accuracy
    • Handmade Symbols: 29.4% accuracy
  • Thicker Conditions:
    • Printed on Paper Media: 88.2% accuracy
    • Handmade Symbols: 95.6% accuracy