EcoScreen AI

• 

  Application with Chemical sunscreen

• 

  Application with Biodegradable sunscreen

• 

  Application starting page

Inspiration

As someone who uses sunscreen often, finding out about the damage it caused to aquatic and marine life shocked me. I mean, it's just sunscreen after all, right? After learning about its impacts, I immediately switched to biodegradable sunscreen, and I hope this application will show others the impacts of their throwaway sunscreen choice.

What it does

The application accepts a picture of the sunscreen's ingredients label, extracts the text from the back, analyses it and submits a concise analysis to the user.

How I built it

Use the OCR.space API for Optical Character Recognition (OCR), I was able to extract the text from the image. Using an API allowed for faster and more robust OCR, while decreasing client-side requirements. In addition, I used Huggingface's API and the Qwen2 model to analyze the ingredients retrieved by the OCR API. This information was then conveyed to the streamlit frontend through a FastAPI, and finally displayed to the user.

Accomplishments that I'm proud of

I am proud of successfully deploying the web-app. It is now available for anybody to use (including you).

What's next for EcoScreen AI

I would like to incorporate more robust OCR techniques, to make it easier for the application to be able to read the ingredients in poor lighting or from odd angles. In addition, I would like to add more information to the application to raise awareness.

Built With

• fastapi
• gpt
• huggingface
• ocr
• python
• streamlit
• threads