SkinTel

Intelligence for your skin. Turn your daily routine into data-driven insights: analyzing your skin, tracking products, and delivering secure, personalized recommendations you can trust.

Comment

Inspiration

Daily skin care routines are essential for many women, but with different skin types and a large variety of products, it can be hard to get started. We want to make it easier for women and others to select skin care products appropriate for their skin (since selecting the wrong ones can have disastrous effects!) where good results will also encourage them to stay with their routine.

What it does

Our app analyzes a person's face locally and then it delete the personally-identifiable photo but it retains skin insights like dryness, oil levels and acne. Then, it analyzes product ingredients and suggests suitable skincare items. We also used the blockchain so that users could anonymously purchase items so they don't have to reveal their private routines and so that they can conveniently use cryptocurrency (or dollars) to purchase products.

How we built it

We used Figma to draw out and design the entire UI/UX, Github for verision control, and React + Next.js with Tailwind CSS for frontend. We additionally used PostgreSQL for the database, and we used FastAPI for the backend.

Challenges we ran into

We were troubled by text detection because there were many hallucinations and typos whenever the ML model analyzed the photo. Therefore, we had to try different strategies to improve accuracy like creating a dictionary of common skincare product ingredients and using a spellchecker library to try to correct minor errors. We ran low on time but we found that a solution using Gemini worked much better but we did not have time to implement this. We also did not have much experience with and Tailwind CSS, so it took us longer than expected to implement pages like the login screen.

Accomplishments that we're proud of

We are proud of our UI because one of our members drew a lot of great cute pictures for our app which made it look so much better! We are also proud of being able to analyze photos to show insights about the user's skin and of being able to incorporate the blockchain and Solana into the app.

What we learned

We learned that we can utilize each other's strengths by dividing the project up into suitable tasks for each person which made the workflow quicker. We also learned more about using AI to improve our productivity and about creating a frontend based off a Figma design.

What's next for SkinTel

With SkinTel, we want to retrieve more product data since there are so many options to choose from like less common foreign brands and some of our images are currently placeholders. We also want to help users create a personalized routine so that they can care for their skin consistently. Furthermore, we would like to look into web deployment by purchasing a domain link and hosting and we are considering expanding SkinTel into an app.

Built With

Share this project:

Updates