🧠 Machine Learning - Healing Bloom Backend

This directory contains the trained Machine Learning models used in the Healing Bloom backend system. These models power core functionalities such as skin disease detection, product recommendation, and acne condition analysis.

📦 Model Overview

1. Skin Disease Identification

Models implemented for this task:

• 🧠 ResNet50
• ⚡ EfficientNet
• 🧪 Custom CNN
• 🌱 DenseNet

These models classify user-uploaded images into various skin disease categories with high accuracy and robustness.

---

2. Product Recommendation

Utilizes the outputs of the disease identification models and additional skin features (e.g., type, condition, sensitivity) to recommend personalized skincare products.

---

3. Acne Condition Detection

Model used:

• 🔍 EfficientNet

This model detects acne levels, including type and spread, using face images. Outputs are then mapped to product suggestions and care routines.

---

📁 Files (not included in this repo)

Due to GitHub's 100MB file limit, the following model files are excluded from this repository:

• final_model.keras
• Other .h5 or .keras files exceeding size limits

You can obtain the model files via:

• Google Drive / S3 (Link here when available)
• Git LFS (if re-enabled)

---

🚀 Usage

These models are loaded into the backend using TensorFlow or Keras and are used in FastAPI endpoints for:

• Disease prediction
• Feature-based product recommendation
• Face color and acne evaluation

Ensure you have the following installed:

pip install tensorflow keras opencv-python pillow