SustainAR

• 

  SustainAR Logo

Inspiration

Lots of the companies spoke out on making software for bettering ourselves and the planet! Thats really cool.

What it does

When the camera is pointed at a object it is able to recognise what it is and return information on the object carbon footprint

How we built it

• Trained a ML tensorflow object labelling algorithm against our set of images
• Used a sample AR app as the foundation of our project
• Plugged in our ML object labelling store into the application
• Made sure that when a object is recognised the correct object and carbon footprint data is returned

Challenges we ran into

• Training the ML algorithm was very tedious and time consuming
• Trying to record a demo video at 3am on halloween weekend
• Building the initial foundation project against our physical devices

Accomplishments that we're proud of

• Finalising a product which utilises both AR and ML tech in a really cool way
• Using and training ML algorithms and having as new found appreciation for them and their uses
• First time writing a project for iOS devices

What we learned

• How long it takes to train a ML algorithm even if its a very small dataset
• swift and objective-c are 2 different things...
• Theres no better way to spend halloween than at a hackathon :P

What's next for SustainAR

• Although continuing on the exact path may not be the future. We did find that using AR and ML is a super cool combo. Hani suggested an idea of building a app which recognises a nespresso pod and gives you information on what kind of coffee it'll make you for his office.

Built With

• ar
• firebase
• ml
• swift
• tensorflow