AstroAnnotate | Devpost

AI Labeling
Home

Inspiration

Our inspiration for this project was to bring pregnant women and those who cannot traditionally work so they could earn money.

What it does

Our web application combines AI apps and uses them to analyze an image with words. A human checks this and confirms if the AI was correct or not as well as a comment as to why/what went wrong. This is then used to determine the correctness of the chosen AI model and pays the user based on their experience accuracy score.

How we built it

We used JavaScript, HTML, and CSS to build the app's front end, utilizing Figma, ProCreate, and Google Fonts to model and create the pages. We also utilized Flask to create a RestAPI connecting our Python backend to our JavaScript. SQLite is used to connect our images and give them their attributes.

Challenges we ran into

Being half of the team's first-time hacking, we ran into getting started with balancing the team's skills and running into version control issues under time constraints. Our project ideas were also very flexible as when we were learning about the prompts we had to change our scope many times.

What's next for AstroAnnotate

This application, while furthering the field of data labeling, also serves as a technically scalable proof-of-concept for ideas that may come afterward. In the future, we hope to add security clearance checking - though the gamified system was created for fun and ease, it also acts as an efficient way to distribute data labeling tasks.