Project Info

AI Skincare Consultant

Devpost

Inspiration

Skincare is hard. Within the ever-expanding beauty industry, it is difficult to keep track of various products and ingredients that are safe to use. For those who are just starting, it can be even harder to navigate the plethora of options and understand what truly benefits their skin. Additionally, personal skin concerns can vary greatly, making it challenging to find tailored solutions without professional guidance. FaceAI Skincare Assistant was inspired by the need to democratize skincare knowledge, making personalized skincare accessible to everyone through the power of artificial intelligence.

What it does

AI Skincare Consultant is a web application designed to simplify and personalize the skincare journey. By uploading a selfie, users can receive an analysis of their skin, identifying issues such as acne, oiliness, pigmentation, and wrinkles. The application then generates personalized skincare routines tailored to the user's unique skin profile, complete with product recommendations and detailed ingredient insights. Additionally, users can interact with the AI assistant to refine their preferences and gain in-depth knowledge about skincare ingredients, learning about the ingredients' usage and safety, ensuring informed and effective skincare choices.

How we built it

The frontend is built with React while OpenCV.js handles image analysis to detect various skin conditions. The backend uses Firebase Cloud Functions and the Gemini 1.5 Flash model to process data and generate personalized skincare recommendations. Vertex AI is utilized for advanced machine learning capabilities, ensuring accurate and reliable analyses. For ingredient data, we use Cosmify's api, which has comprehensive data for various ingredients used in skincare products.

Challenges we ran into

Image Analysis: It was difficult to get image analysis that was accurate. Skin conditions can often be difficult to identify and it was clear that there were still inaccuracies in our evaluations. As such, we had users confirm the conclusions that the AI drew in case they were incorrect. Personalization Complexity: Creatin personalized skincare routines involves considering a multitude of factors, including skin type, concerns, sensitivities, and user preferences, which added complexity to the recommendations.

Accomplishments we're proud of

We are proud that we were able to leverage various technologies to create a web app that could potentially help many people in their skincare journeys.

What we learned

Throughout our development process, we learned much about Gemini's capabilities and weaknesses. Although it is certainly powerful, there are aspects in image analysis where it fails, which was important to our project. Also, we were able to learn more about the development process, constantly finding ways to improve on our ideas.

What's next

I would like to continue in the development of the skincare consultant, implementing various new features that can help users including Routine Progress Tracking Expanded Skin Condition Detection Direct Skincare Product Integration Enhanced Personalization Community Features

Analysis

Compare with all teams

View

Metric

Figures cover GitHub contributors during the hackathon window. A co-authored commit counts in full for each author, so per-member totals add up to more than the whole-team figures.

Technology

Found in codeClaimed only
  • CSSIn code
  • ExpressIn code
  • FirebaseIn code
  • HTMLIn code
  • JavaScriptIn code
  • ReactIn code
  • Google GeminiClaimed

6 of 7 appear in the indexed code. 1 claimed on Devpost could not be matched to code, which may simply mean the tool leaves no trace in the repository.

AI coding agents

No AI coding agent signals were found in this repository.

Detected from committed agent config files and commit authorship. Absence of a signal is not proof an agent was unused.

Codebase size

Source size

61 KB

Source files

14

Counts recognized source files only; vendored directories, binaries and lockfiles are excluded, so this is smaller than the repository on disk.

0 stars