Recycraft

Bangkit 2022 Product Based Capstone

Team ID : C22-PS285

Members :

• (M2007F0776) - Wahyu Adi Nugroho - Universitas Dian Nuswantoro
• (M2007F0777) - Raphael Adhimas Aryandanu Santoso - Universitas Dian Nuswantoro
• (A2007F0724) - Alfonda Steven Wahyudi - Universitas Dian Nuswantoro
• (A2009F0915) - Firstiannisa Rizki - Universitas Gunadarma
• (C2009F0934) - Zekri Fitra Ramadhan - Universitas Gunadarma
• (C7009F0926) - Muhammad Rafi Ramadhan - Universitas Gunadarma

Recyraft-Machine Learning

This project is our final project for Google Bangkit Academy 2022.

Recyraft-Android : RecyraftApp

Recyraft-Cloud : RecyraftCloud

Machine Learning : Building two models that include : scraps type classification and scraps classification detection. Build process using baseline experiment, dropout, flatten. In this project, we use simple Convolutional Neural Network and pre-trained model or transfer learning by Inceptionv3 and Xception. The model was saved with model.tflite and chosen by the best model for deployment.

Machine Learning Task :

1. Scraps Type Classification (Binary Classification)
2. Scraps Classification & Detection (Multiple Classification)

Datasets (Kaggle and Github) :

1. Recycleable or Organic :

• Waste Classification Data

2. Scraps with 8 class ( Bottle-Plastic, Can, Cardboard, Glass-Plastic, Paper, Plastic, Spoon-Plastic, and Styrofoam) :

• Garbage Classification - 12 Classes
• Most Common Recyclable and Non-Recyclable Objects
• TrashBox
• Crapping Image on Google Search (Manually)

Features :

1. EDA (Exploratory Data Analysis) for Image Classification
2. Preprocessing Data and Image
3. Image Augmentation
4. Simple CNN
5. InceptionV3
6. Xception
7. TFLite x labels

Prerequisites :

1. Anaconda (Jupyter Notebook) or Google Colab
2. Python version 3.7 or above
3. Tensorflow 2.8 or latest version
4. Tensorflow Lite
5. Keras API
6. Kaggle API Token → Generate

Documentation :

1. Import Library and Preparing Dataset
2. Splitting and Checking Dataset
3. Preprocessing Dataset and Perform Data Augmentation
4. Make ML Model, Build and Training Dataset
5. Model Evaluation
6. Create Prediction Data
7. Get Result Prediction
8. Saved Model and Convert to TFLite Model

References (Paper/Journal/Article) :

1. https://iopscience.iop.org/article/10.1088/1755-1315/775/1/012010/meta
2. https://www.mdpi.com/2313-4321/7/1/9
3. https://www.scirp.org/pdf/jcc_2021011322480971.pdf
4. https://ieeexplore.ieee.org/abstract/document/9395916