Hi, this project is called WasteCreative. It is an app which its main feature is to help you utilize the waste on your house, mainly household waste. This project is part of my Bangkit Journey as the capstone project we I take the role of machine learning engineer. The repo can be found here https://github.com/Afrizal2609/WasteCreative
Dataset
The dataset itself consist of scraped images of 12 objects. We gather the image ourself using a tool.
Tools & Technology (ML Part)
There are some tools and tech we use on this project.
- Google Image Scraper: this is used to gather images from Google for certain objects. The tool can be found here https://github.com/ahmadnurokhim/Google-Image-Scraper
- LabelImg: this tool is used to annotate the gathered images for certain objects. The tool can be found here https://github.com/HumanSignal/labelImg
- Roboflow: this tool is used to do the data augmentation. The tool can be found here https://roboflow.com/
- Tensorflow Object Detection API: this API is used to use the retrain dataset to match our objects. We choose to retrain the SSD MobileNet V2 FPNLite 320×320 since we want to deploy the model on mobile platform. The API can be found here https://github.com/tensorflow/models/tree/master/research/object_detection
After finished the training, we convert the model into TFLite format so it can be deployed on mobile.
Gallery
