Wild-In

• 

  Main interface for found ingredients

Inspiration

The intersection between adventure and home cooking.

What it does

We use image recognition to scan your surroundings for edible plants. Then compile a recipe book using those ingredients. The ingredients and recipes can be found in a gamified and clean UI that allows users to venture into the wild sin a whole new way!

How we built it

1. Retrained YOLOv5s to recognize edible plants in the wild.
2. Used ChatGPT to help with recipe generation.

Documentation

link

Challenges we ran into

1. Training a medium scale model within the allotted timeframe.
2. Finding a relevant dataset.
3. Integrating flask and react for a seamless experience.

Accomplishments that we're proud of

1. Effective retraining of YOLO for specific use cases beating out chatGPT in some cases.
2. Clean and thematic GUI.

What we learned

How to integrate a full application in a small team.

What's next for Wild-In

1. Expanding the dictionary of ingredients that Wild-In can recognize.
2. Implementing a feature to compare ingredients found in the wild with common household ingredients.

Built With

• flask
• openai
• python
• pytorch
• react
• yolo