LOW TAIPER FILTER

• 

  ui

• 

  ui

• 

  Edited image using our agent chain

• 

  edited image with our software

Inspiration

Snapchat, Adobe Lightroom and Instagram

What it does

It generates image filtering code and applies it to your input image.

How we built it

We used Python as a backend calling OpenAI's API to generate Python code using the Pillow module to manipulate the image, we then display the filtering code and the image of the Flask + HTML + CSS frontend.

Challenges we ran into

Originally tried to use Next.JS but couldn't integrate it correctly with the backend. We also had to scope down our AI agents to 2 agents rather than 11 that we manged to get running in OpenAI playground.

Accomplishments that we're proud of

We have a working prototype and have used it to generate lots of filters!

What we learned

None of us had ever used the OpenAI API so we had to learn quickly. We also learned the limits of the AI. Some of us also learnt how to scope a software project as we are engineers and only done hardware hackathons before.

What's next for LOW TAIPER FILTER

We hope to integrate the better AI agent system into our code to improve the filtering algorithms. We would also want to implement it in industrial as generating image analysis code can be very useful in medical and lab environments. Perhaps let radiologists use it to enhance scans in different ways to make them easy to analyse. Could also have hobbyist uses to provide no-code image analysis in areas that typically use code heavy statistics such as satellite image analysis.

Built With

• flask
• html5
• openai
• python