Pixels2GenAI

An open source educational platform that creates a comprehensive pathway into AI-driven generative art, bridging mathematical and visual foundations to modern creative AI techniques. This 15-module curriculum takes learners from fundamental pixel manipulation and NumPy operations through advanced generative models, neural networks, and real-time interactive systems.

Showcasing exercises from across all modules

Project at a Glance

Educational Pathway

Creating an approachable journey into AI-driven generative art that connects visual intuition with modern machine learning practice through 15 progressive modules.

Designed for Learners

Materials welcome semi-beginners through semi-experienced programmers, with optional guidance for newcomers willing to self-study foundational topics.

Theory Meets Practice

Each module balances mathematical ideas, NumPy techniques, and creative coding projects so learners see how concepts translate into visuals and AI applications.

Progressive Curriculum

Lessons build sequentially from image fundamentals through fractals, simulations, and generative AI, allowing confidence to grow alongside complexity.

Creative Community Impact

Supports programming teachers, self-learners, artists, and data scientists who want memorable exercises for classes, portfolios, or passion projects.

Use Cases & Audiences

Ideal for course builders, independent learners, curious engineers, and creatives exploring AI-enhanced artistry across classrooms and studios.

Repository

The source code is available on GitHub:

https://github.com/burakkagann/Pixels2GenAI

Clone the repository:

git clone https://github.com/burakkagann/Pixels2GenAI.git
cd Pixels2GenAI

Installation

System Requirements:

• Python 3.11.9 (recommended)

• For neural network modules (7+): NVIDIA GPU recommended but not required

• For diffusion models (Module 12): 8GB RAM minimum, GPU strongly recommended

Option 1: Using pyproject.toml (Recommended)

# Core dependencies (Modules 0-6)
pip install .

# With machine learning packages (Modules 7-13)
pip install .[ml]

# All optional dependencies
pip install .[all]

Option 2: Using requirements.txt

pip install -r requirements.txt

Learning Modules

Creative Coding Foundations

Modules 0-6 · ~80 exercises · Start here if new to creative coding

Module 0: Foundations & Definitions

Setting the conceptual and technical groundwork for generative art and AI.

• 0.1 - What Is Generative Art
• 0.2 - Defining AI ML Algorithms
• 0.4 - Setup Environment

Module 1: Pixel Fundamentals

Understanding images at the atomic level through color theory and manipulation patterns.

1.1 - Grayscale & Color Basics

• 1.1.1 - Color Basics

1.1.2 - Color Theory Spaces

1.2 - Pixel Manipulation Patterns

• 1.2.1 - Random Patterns
• 1.2.2 - Cellular Automata

1.2.3 - Reaction Diffusion

1.3 - Structured Compositions

1.3.3 - Truchet Tiles 1.3.4 - Wang Tiles

Module 2: Geometry & Mathematics

Mathematical foundations for generative art through shapes, coordinates, and mathematical patterns.

2.1 - Basic Shapes & Primitives

• 2.1.1 - Lines
• 2.1.2 - Triangles
• 2.1.3 - Circles
• 2.1.4 - Stars

2.1.5 - Polygons & Polyhedra

2.2 - Coordinate Systems & Fields

• 2.2.1 - Gradient
• 2.2.2 - Spiral
• 2.2.3 - Vector Fields
• 2.2.4 - Distance Fields

2.3 - Mathematical Art

• 2.3.2 - Rose Curves
• 2.3.3 - Harmonograph Simulation

2.3.1 - Lissajous Curves 2.3.4 - Strange Attractors

Module 3: Transformations & Effects

Manipulating visual data through geometric transformations, masking, and artistic filters.

3.1 - Geometric Transformations

• 3.1.1 - Rotation
• 3.1.2 - Affine Transformations
• 3.1.3 - Nonlinear Distortions
• 3.1.4 - Kaleidoscope Effects

3.2 - Masking & Compositing

3.2.1 - Mask 3.2.2 - Meme Generator 3.2.3 - Shadow 3.2.4 - Blend Modes

3.3 - Artistic Filters

• 3.3.1 - Warhol
• 3.3.3 - Hexpanda
• 3.3.5 - Delaunay Triangulation

3.3.2 - Puzzle (Array Concatenation) 3.3.4 - Voronoi Diagrams

3.4 - Signal Processing

• 3.4.1 - Convolution

3.4.2 - Edge Detection (Sobel Operator) 3.4.3 - Contour Lines 3.4.4 - Fourier Art

Module 4: Fractals & Recursion

Self-similarity and infinite complexity through classical fractals, natural patterns, and L-systems.

4.1 - Classical Fractals

• 4.1.1 - Fractal Square
• 4.1.2 - Dragon Curve
• 4.1.3 - Mandelbrot

4.1.4 - Julia Sets 4.1.5 - Sierpinski

4.2 - Natural Fractals

• 4.2.1 - Fractal Trees

4.2.2 - Lightning Bolts 4.2.3 - Fractal Landscapes 4.2.4 - Diffusion Limited Aggregation

4.3 - L-Systems

4.3.1 - Plant Generation 4.3.2 - Koch Snowflake 4.3.3 - Penrose Tiling

Module 5: Simulation & Emergent Behavior

Complex systems from simple rules: particle systems, flocking behavior, and physics simulations.

5.1 - Particle Systems

• 5.1.1 - Sand

5.1.2 - Vortex 5.1.3 - Fireworks Simulation 5.1.4 - Fluid Simulation

5.2 - Flocking & Swarms

• 5.2.1 - Boids

5.2.2 - Fish Schooling 5.2.3 - Ant Colony Optimization

5.3 - Physics Simulations

• 5.3.3 - Double Pendulum Chaos

5.3.1 - Bouncing Ball Animation 5.3.2 - N-Body Planet Simulation 5.3.4 - Cloth Rope Simulation 5.3.5 - Magnetic Field Visualization

5.4 - Growth & Morphogenesis

5.4.1 - Eden Growth Model 5.4.2 - Differential Growth 5.4.3 - Space Colonization Algorithm 5.4.4 - Turing Patterns

Module 6: Noise & Procedural Generation

Controlled randomness for natural effects: noise functions, terrain, textures, and wave patterns.

6.1 - Noise Functions

• 6.1.1 - Perlin Noise

6.1.2 - Simplex Noise 6.1.3 - Worley Noise 6.1.4 - Colored Noise

6.2 - Terrain Generation

6.2.1 - Height Maps 6.2.2 - Erosion Simulation 6.2.3 - Cave Generation 6.2.4 - Island Generation

6.3 - Texture Synthesis

6.3.1 - Marble Wood Textures 6.3.2 - Cloud Generation 6.3.3 - Abstract Patterns 6.3.4 - Procedural Materials

6.4 - Wave & Interference Patterns

6.4.1 - Moire Patterns 6.4.2 - Wave Interference 6.4.3 - Cymatics Visualization

ML & Animation

Modules 7-9 · ~36 exercises · Machine learning and motion

Module 7: Classical Machine Learning

Traditional ML for creative applications: clustering, classification, and statistical methods.

7.1 - Clustering & Segmentation

7.1.1 - KMeans Clustering 7.1.2 - Meanshift Segmentation 7.1.3 - DBSCAN Pattern Detection

7.2 - Classification & Recognition

7.2.1 - Decision Tree Classifier 7.2.2 - Random Forests 7.2.3 - SVM Style Detection

7.3 - Dimensionality Reduction

7.3.1 - PCA Color Palette 7.3.2 - t-SNE Visualization 7.3.3 - UMAP Visualizations

7.4 - Statistical Methods

7.4.1 - Monte Carlo Sampling 7.4.2 - Markov Chains 7.4.3 - Hidden Markov Models

Module 8: Animation & Time

Adding the fourth dimension: animation fundamentals, organic motion, and cinematic effects.

8.1 - Animation Fundamentals

8.1.1 - Image Transformations 8.1.2 - Easing Functions 8.1.3 - Interpolation Techniques 8.1.4 - Sprite Sheets

8.2 - Organic Motion

• 8.2.2 - Infinite Blossom

8.2.1 - Flower Assembly 8.2.3 - Walk Cycles 8.2.4 - Breathing Pulsing

8.3 - Cinematic Effects

• 8.3.1 - Star Wars Titles
• 8.3.2 - Thank You

8.3.3 - Particle Text Reveals 8.3.4 - Morphing Transitions

8.4 - Generative Animation

8.4.1 - Music Visualization 8.4.3 - Animated Fractals 8.4.2 - Data Driven Animation

Module 9: Introduction to Neural Networks

Bridge to modern AI: neural network fundamentals, architectures, and training dynamics.

9.1 - Neural Network Fundamentals

• 9.1.1 - Perceptron Scratch
• 9.1.2 - Backpropagation Visualization
• 9.1.3 - Activation Functions Art

9.2 - Network Architectures

9.2.1 - Feedforward Networks 9.2.2 - Convolutional Networks Visualization 9.2.3 - Recurrent Networks for Sequences

9.3 - Training Dynamics

9.3.1 - Loss Landscape Visualization 9.3.2 - Gradient Descent Animation 9.3.3 - Overfitting Underfitting Demos

9.4 - Feature Visualization

9.4.1 - DeepDream Implementation 9.4.2 - Feature Map Art 9.4.3 - Network Attention Visualization

Real-Time & AI Integration

Modules 10-13 · ~48 exercises · TouchDesigner and generative AI

Module 10: TouchDesigner Fundamentals

Real-time visual programming: TD environment, NumPy integration, and interactive controls.

10.1 - TD Environment & Workflow

• 10.1.1 - Node Networks

10.1.2 - Python Integration Basics 10.1.3 - Performance Monitoring

10.2 - Recreating Static Exercises

10.2.1 - Core Exercises Realtime 10.2.2 - Boids Flocking in TouchDesigner 10.2.3 - Planet Simulation TD 10.2.4 - Fractals Realtime

10.3 - NumPy to TD Pipeline

10.3.1 - Script Operators 10.3.2 - Array Processing 10.3.3 - Custom Components

10.4 - Interactive Controls

10.4.1 - UI Building 10.4.2 - Parameter Mapping 10.4.3 - Preset Systems

Module 11: Interactive Systems

Sensors and real-time response: input devices, computer vision, and physical computing.

11.1 - Input Devices

11.1.1 - Webcam Processing 11.1.2 - Audio Reactivity 11.1.3 - MIDI OSC Control 11.1.4 - Kinect Leap Motion

11.2 - Computer Vision in TD

• 11.2.3 - Face Detection

11.2.1 - Motion Detection 11.2.2 - Blob Tracking 11.2.4 - Optical Flow

11.3 - Physical Computing

11.3.1 - Arduino Integration 11.3.2 - DMX Lighting Control 11.3.3 - Projection Mapping Basics

11.4 - Network Communication

11.4.1 - Multi Machine Setups 11.4.2 - WebSocket WebRTC 11.4.3 - Remote Control Interfaces

Module 12: Generative AI Models

Modern generative techniques: GANs, VAEs, diffusion models, and language models for art.

12.1 - Generative Adversarial Networks

• 12.1.1 - GAN Architecture
• 12.1.2 - DCGAN Art
• 12.1.3 - StyleGAN Exploration
• 12.1.4 - Pix2Pix Applications

12.2 - Variational Autoencoders

• 12.2.1 - Latent Space Exploration
• 12.2.2 - Interpolation Animations
• 12.2.3 - Conditional VAEs

12.3 - Diffusion Models

• 12.3.1 - DDPM Basics
• 12.3.2 - ControlNet Guided Generation

12.4 - Bridging Paradigms

12.4.1 - Neural Style Transfer 12.4.2 - VQ-VAE and VQ-GAN

12.5 - Personalization & Efficiency

• 12.5.1 - DreamBooth Personalization

12.6 - Transformer Generation

12.6.1 - Taming Transformers 12.6.2 - Diffusion Transformer (DiT)

12.7 - Modern Frontiers

• 12.7.1 - Flow Matching

Module 13: AI + TouchDesigner Integration

Combining AI with real-time systems: ML models in TD, real-time effects, and hybrid pipelines.

13.1 - ML Models in TD

13.1.1 - MediaPipe Integration 13.1.2 - RunwayML Bridge 13.1.3 - ONNX Runtime

13.2 - Real-time AI Effects

13.2.1 - Style Transfer Live 13.2.2 - Realtime Segmentation 13.2.3 - Pose Driven Effects

13.3 - Generative Models Live

13.3.1 - GAN Inference Optimization 13.3.2 - Latent Space Navigation UI 13.3.3 - Model Switching Systems

13.4 - Hybrid Pipelines

13.4.1 - Preprocessing TD 13.4.2 - Python ML Processing 13.4.3 - Post Processing Chains

Data & Capstone

Modules 14-15 · Final projects

Module 14: Data as Material

Information visualization and sonification: data sources, visualization techniques, and physical sculptures.

14.1 - Data Sources

14.1.1 - APIs and Data Scraping 14.1.2 - Sensor Networks 14.1.3 - Social Media Streams 14.1.4 - Environmental Data

14.2 - Visualization Techniques

14.2.1 - Network Graphs 14.2.2 - Flow Visualization 14.2.3 - Multidimensional Scaling 14.2.4 - Time Series Art

14.3 - Sonification

14.3.1 - Data Sound Mapping 14.3.2 - Granular Synthesis 14.3.3 - Rhythmic Patterns

14.4 - Physical Data Sculptures

14.4.1 - 3D Printing Preparation 14.4.2 - Laser Cutting Patterns 14.4.3 - CNC Toolpaths

Module 15: Capstone Project - Eternal Flow

Synthesis of all learned concepts: StyleGAN-based evolving Ebru marbling artwork for projection display.

License

This work is licensed under the MIT License.

• Burak Kağan Yılmazer (2025) — burak.kagan@protonmail.com

• Dr. Kristian Rother (2024) — kristian.rother@posteo.de

See LICENSE for full license terms.

Built With

This curriculum is built with:

Core Computing: NumPy, SciPy, pandas

Image Processing: Pillow, OpenCV, ImageIO

Machine Learning: scikit-learn, PyTorch, TensorFlow

Visualization: matplotlib, Jupyter

Real-time Systems: TouchDesigner

For academic citations and detailed references, see individual module documentation.

Inspiration

• Generative Art: A Practical Guide by Matt Pearson

• Processing: A Programming Handbook by Casey Reas & Ben Fry