This is repository for Billy. Billy is a prototype that we build for our participation in AWS Hackday 2019. Billy has been selected as the top 5 teams from Singapore and was invited to pitch this projects at the national Demo Day in Singapore (on 29 March 2019).
Our team member:
- Daniel
- Agus Ekanurdi
- Muhammad Khairul Syamil Hashim
Billy is an mobile or IoT application that is used in waste management. This application can do object classification so that user is able to know which correct recycle bin to throw it. In this prototype, we do classification for 6 different type of objects. Those are metal, paper, cardboard, plastic, glass and trash.
We build this application on AWS infrastructure and the main service we use are AWS Rekognition and AWS SageMaker. In this case, AWS Rekognition does object detections and returns possible labels (and its confident levels). Those labels are inputted into AWS SageMaker to predict type of object.
For training purpose in AWS SageMaker, we are use trash images dataset from https://github.com/garythung/trashnet
Folder description:
a. backend
It contains our code for managing and orchestrating request(s) from mobile application/IoT devices.
b. dataset
It contains trash images dataset for training and testing purposes.
c. documents
It contains our documentations including our executive summary and demo video.
d. frontend
It contains our code for our prototype of mobile application.
e. lambda
It contains our code in AWS Lambda to orchestrator process to Amazon Rekognition and Amazon SageMaker.
f. training
It contains our code in Jupiter Notebook to build and train our model in Amazon SageMaker