FreshScan

• 

  3D Model Asset

Inspiration

FreshScan Idea

What it does

Through FreshScan, farmers can complete their profile based on their product. This data is stored in our system and generates a QR code which goes on the produce, which is then shipped to stores. Upon scanning the QR code, customers can see a 3D Asset and Info page which contains a YouTube link about that particular crop. Users can watch the educational YouTube video and then take a quiz. Lastly, we have a reward system, through which users get more points the more videos they watch.

How we built it

In order to build our application, we used:

• Microsoft Azure to build our entire backend
• *EchoAR to generate real-time
• React native to build our frontend app
• Python to retrieve transcription of youtube videos and convert to quizzes

Challenges we ran into

• We struggled with getting YouTube transcription from Python, and generating a quiz in realtime, but we were able to pull it off
• Struggled with creating echoAR models

Accomplishments that we're proud of

• Proud of creating a coherent system that functions smoothly across the frontend and backend interfaces
• Proud of creating an EchoAR model for fruits such as apple, banana, blueberry in realtime and integrating with our app

What we learned

• Learned how to use Microsoft Azure and cosmosDB to store data, and access through the frontend

What's next for FreshScan

• Next up, we would like to spread our technology across farmers, and get many farmers to create their profiles
• Furthermore, we would like to integrate our QR code system with existing sticker manufacturing companies

Built With

• azure
• echoar
• python
• pytorch
• react
• react-native
• tensorflow