3D Deep Learning Network Visualization

This project is a 3D visualization tool for deep learning network architectures using OpenGL and WinAPI. The application renders predefined architectures (such as AlexNet, VGG16, and ResNet18) or allows users to define their own custom neural networks.

Features

• Render convolutional layers as 3D boxes.
• Render fully connected layers as spheres with interconnections.
• Interactive camera controls with mouse rotation.
• Predefined networks: AlexNet, VGG16, ResNet18.
• Custom network setup via console input.

Dependencies

• GCC (MinGW recommended for Windows)
• OpenGL (-lopengl32)
• GLU (-lglu32)
• Math library (-lm)
• Windows API (-mwindows)

Installation & Execution

1. Install MinGW (if not installed)

Ensure that gcc is available in your system's PATH. You can install MinGW from:

• https://www.mingw-w64.org/

2. Compile the Program

Run the following command in the terminal:

gcc main.c -o deep3d -mwindows -lopengl32 -lglu32 -lm

This will produce an executable named deep3d.exe.

3. Run the Application

Execute the generated binary:

./deep3d.exe

Controls

• Mouse Drag: Rotate the view.
• Left Click: Hold to enable rotation.

Code Structure

1. Core Functions

• SetupConsole(): Initializes the console for debugging.
• SetupNetwork(): Loads predefined or custom network architecture.
• InitOpenGL(): Configures OpenGL for rendering.
• RenderScene(): Draws the neural network.

2. Network Representation

• Boxes represent convolutional layers.
• Spheres represent fully connected neurons.
• Arrows indicate layer connections.

3. Predefined Architectures

• SetupAlexNet()
• SetupVGG16()
• SetupResNet18()

Custom Network Setup

1. Select Custom Network in the menu.
2. Enter the number of layers.
3. Define layer type (Box for Conv layers, FC for fully connected layers).
4. Specify layer dimensions and color.

Example Output

A rendered AlexNet architecture would look like:

[Input] → [Conv1] → [Conv2] → [Conv3] → [Conv4] → [Conv5] → [FC6] → [FC7] → [FC8]

Troubleshooting

• If the window does not appear, ensure OpenGL drivers are installed.
• If compilation fails, check that MinGW is correctly installed and in the system PATH.

License

This project is open-source and available for modification and distribution.

---

Developed for 3D visualization of neural networks using OpenGL and WinAPI.