fAIshion

Inspiration

We were brainstorming ideas and came up with this one the most. We also thought we had a good amount of knowledge in fashion.

What it does

It assists you with choosing clothing items that pair well with your physical features.

How we built it

We used Python's several ML/AI libraries (mediapipe, sklearn, CV2, etc.) and used AI assistance (as described in our GitHub README) to help us with

Challenges we ran into

We wanted to do a User Interface, but decided to focus on trying to implement functions first by reading documentation, and then used AI assistance to develop a drag-and-drop pipeline.

Accomplishments that we're proud of

Understanding of several ML/AI libraries and our presentation.

What we learned

We learned how to leverage LLMs effectively and about several Python libraries, in particular, MediaPipe.

What's next for fAIshion

Fine-tuning an LLM as opposed to few-shot prompting. Graphs/maps in your personal wardrobe that will pair clothing items. For accurate skin tone prediction, we plan to use a face input for a clearer color. Selecting style types.

Built With

• mediapipe
• numpy
• openai
• python
• sklearn