CelebMax

CelebMax is a web application that allows users to upload a photo of themselves and receive personalized hairstyle recommendations based on celebrity lookalikes. Our system uses a trained AI model to compare the user’s image against a dataset of 25,000 celebrity images, selecting the top 5 matches. Users can then choose one of these celebrities and apply the celebrity’s hairstyle onto their own photo.

Created By:

• Jacob Hoppenstedt
• Daniel Dovale
• Rohith Vellore
• Veeraj Reddy Talasani

---

Table of Contents

• Features
• Getting Started
  • Prerequisites
  • Installation
  • Setting Up External Files
• Docker Compose Setup
• Contributing
• License

---

Features

• Celebrity/Model Matching:

  • Users upload a photo.
  • The backend calls an AI model that searches through a dataset of 25,000 celebrity images.
  • The system picks the top 5 most similar faces based on facial similarity.

• Hairstyle & Outfit Recommendations:

  • Once a celebrity match is selected, users receive hairstyle recommendations.
  • Integration with HairFastGAN-AI allows users to overlay the celebrity’s hair onto their own image.

• User Authentication:

  • Secure signup and login functionalities with basic authentication.

• Modular Architecture:

  • A Go backend that handles API requests and user authentication.
  • A Python-based microservice that performs AI model inference.

• Docker Compose Integration:

  • Use Docker Compose to build and run the entire stack with a single command.

---

Getting Started

Prerequisites

• Docker & Docker Compose:
  Ensure you have Docker installed on your machine. Get Docker

• Go 1.19+

• Python 3.8+ (handled via Docker in our setup)

Installation

1. Clone the Repository:

   git clone https://github.com/jacobhoppenstedt/celebmax.git
   cd celebmax
   

2. Install Go Dependencies:

   go mod tidy

3. Configure Environment Variables: Create a .env file in the root of the project with the following content:

PYTHON_SERVICE_URL=http://localhost:5001

Setting Up External Files

The project requires several large external files (e.g., pre-trained model weights and parts of the 25,000 celebrity dataset). These files are not included in the repository to keep the repository lightweight.

To set up these files:

Download the Files:

Download the required files from our Google Drive folder.

Place the Files:

After downloading, place the files into the python-service directory. The expected structure is:

celebmax/
├── python-service/
│   ├── Haircut/
│   │   └── (files related to HairFastGAN-AI or hairstyle processing)
│   ├── .gitattributes
│   ├── aws_credentials.py
│   ├── clustered_embeddings.npy
│   ├── Dockerfile
│   ├── kmeans_model.pkl
│   ├── main.py
│   ├── requirements.txt
│   └── tests3.py
├── backend/
└── docker-compose.yml

Set up Fal.AI and AWS API Keys

• Obtain Fal.ai API key, create API_KEY.py file and place inside python-service/Haircut folder.
• Obtain AWS key and update values inside the aws_credentials.py

Docker Compose Setup

A docker-compose.yml file is provided in the root of the project to orchestrate the Go backend and the python-service.

Build and Run the Containers

From the project root, run:

docker-compose up --build

This command will build the images (if necessary) and start the services:

• Backend (Go Server): Runs on port 8080.

• Python-service: Runs on port 5001.

Contributing

We welcome contributions! To contribute:

1. Fork the repository.

2. Create a new branch for your feature or bug fix.

3. Follow the coding standards and include tests for any new functionality.

4. Submit a pull request with a clear description of your changes.

License

This project is licensed under the MIT License. See the LICENSE file for details.