📦 AI Packaging Optimizer

An intelligent packaging optimization tool powered by Convolutional Neural Networks (MobileNetV2) and OpenCV. This system automatically analyzes product images, extracts physical dimension features (Color Histograms & HOG), and predicts the most optimal standard box size (e.g., Flat mailer, 6x4x4, 12x12x8) to minimize shipping void space and waste.

🚀 Key Features

• Computer Vision Extraction: Calculates Histogram of Oriented Gradients (HOG) and color properties from raw images.
• Deep Learning Prediction: Uses a fine-tuned MobileNetV2 architecture with custom regression heads.
• Data Pipelines: Batch compresses, cleans, and augments e-commerce dataset JSON/CSV structures (ABO datasets).
• Physical Bounding Math: Computes custom padding offsets and optimal 3D rotational fits for standard packaging.

🛠️ Getting Started

python prepare_data.py
python extract_features.py
python train_model.py
python predict_packaging.py