Structured learning paths for computer vision and AI
Our curriculum breaks down complex topics into focused modules. Each program covers technical foundations, practical implementation, and real project work. You learn from researchers who work with these systems daily, using datasets and tools that reflect current industry practice. Expect detailed explanations, coding assignments, and feedback on your work.
What each module includes
Six core modules, each running four weeks with specific learning objectives
Image fundamentals
4 weeks В· 28 hours
How images work as data structures. Color spaces, pixel manipulation, filtering operations, and edge detection using OpenCV and NumPy.
- Kernel operations and convolution
- Histogram analysis and equalization
- Morphological transformations
- Feature extraction basics
Neural network architecture
4 weeks В· 32 hours
Building and training convolutional networks. Layer design, activation functions, backpropagation, and optimization techniques with PyTorch.
- CNN layer configuration
- Training loop implementation
- Loss function selection
- Regularization methods
Object detection systems
4 weeks В· 30 hours
Detecting and localizing multiple objects in images. YOLO architecture, region proposals, non-maximum suppression, and bounding box regression.
- Anchor box configuration
- IoU calculation and matching
- Multi-scale detection
- Real-time inference optimization
Semantic segmentation
4 weeks В· 30 hours
Pixel-level classification for scene understanding. U-Net and DeepLab architectures, encoder-decoder structures, and segmentation metrics.
- Skip connection implementation
- Atrous convolution techniques
- Dice coefficient optimization
- Multi-class segmentation
Video processing
4 weeks В· 28 hours
Temporal analysis and tracking across frames. Optical flow, motion estimation, and multi-object tracking with Kalman filters.
- Frame differencing methods
- Tracking algorithm comparison
- Temporal coherence maintenance
- Background subtraction techniques
Production deployment
4 weeks В· 26 hours
Taking models from notebooks to production systems. Model quantization, ONNX conversion, API design, and inference optimization strategies.
- Model compression techniques
- Batch processing setup
- GPU memory management
- Performance monitoring
How you progress through the program
Foundation phase
First eight weeks cover image processing fundamentals and basic neural networks. You build simple classifiers and understand how data flows through layers.
Weeks 1-8
Application phase
Next eight weeks focus on detection and segmentation. You work with larger models and learn to handle complex visual tasks with multiple outputs.
Weeks 9-16
Integration phase
Final eight weeks combine video processing with deployment strategies. You optimize models for speed and build complete inference pipelines.
Weeks 17-24
Who teaches these modules
Dr. Linnea Vesterholm
Image processing specialist
Leads the foundation modules covering image fundamentals and feature extraction. Published research on efficient convolution operations and teaches the math behind kernel design.
Siobhan MacLeod
Deep learning architect
Handles network architecture and training modules. Works on medical imaging systems and focuses on practical implementation details that affect model performance.
Annika JГёrgensen
Production systems engineer
Teaches video processing and deployment modules. Builds real-time detection systems for autonomous vehicles and knows optimization techniques that actually matter in production.