Fashion Fusion

Inspiration

We were inspired by Inditex Tech, whose challenge was very interesting.

What it does

Given an image introduced by the user, the system recommends a piece of clothing.

How we built it

Backend and fronted were programmed in Python and JavaScript using both React and FastAPI. The AI part used Pytorch and the database was built in Supabase,

Challenges we ran into

The large computational capacity required, the large amount of time it took to pre-process data and train the algorithm, and the new concepts related to similarity.

Accomplishments that we're proud of

Have learned about full-stack technologies, mathematical concepts applied to similarity estimation in the field of computer vision and some AI related frameworks.

What we learned

We gained valuable insights into effective teamwork and time management. Despite facing time problems, we collaborated and communicate effectively, and prioritize some tasks to meet the deadline.

What's next for Fashion Fusion

In the next phase of our project, we aim to optimise the number of clusters to reduce the data volume and improve the efficiency of our recommendation system. Furthermore, we should explore alternative measures of similarity, such as cosine similarity or Jaccard index. Moreover we plan to improve the user experience by integrating direct links to purchase recommended items from popular online retailers and refine the web design to ensure intuitive navigation.

Built With

• fastapi
• javascript
• python
• pytorch
• react
• supabase