VR_AI_Decision_Support_System

This project uses a fine-tuned VGG16 neural network to classify images into three distinct styles: Minimal, Modern, and Traditional. The code is written in Python, and the trained model is exported as an ONNX file for integration into a Unity project.

How It Works

1. Dataset Structure:
   To train and test the model, the code expects datasets to follow the structure:

• The train folder contains training images divided into three style categories: minimal, modern, and traditional.
• The test folder contains corresponding test images, following the same category structure.
• Note: Ensure the folder structure remains unchanged after downloading the dataset.

Download the dataset from the provided Link and place the DS folder in the same directory as the code. Ensure the folder structure remains unchanged.

2. Training:
   The script trains the VGG16 model using the data in DS/train. The test dataset in DS/test is used for evaluation.

3. Export:
   After training, the model is exported as an ONNX file, making it compatible with Unity projects.

Prerequisites

• Python 3.x
• Required Python packages (exact versions will be specified later).

Steps to Run

1. Download the dataset from Link.
2. Place the DS folder next to the project files.
3. Install the required packages:

pip install -r requirements.txt