Step 1readingRequired360 minutes
Mathematical foundations
Review vectors, matrices, convolution, probability and optimisation concepts used throughout computer vision.
Path builder
Computer Vision Foundations
Add resources, projects, exercises and milestones, then publish the path.
Curriculum
Path steps
Step 2readingRequired420 minutes
Image representation and processing
Understand pixels, colour spaces, filtering, edges, features and geometric transformations.
Step 3exerciseRequired480 minutes
OpenCV fundamentals
Practice loading, transforming, filtering and analysing images with OpenCV.
Step 4readingRequired600 minutes
Convolutional neural networks
Study convolution, pooling, receptive fields, training and common CNN architectures.
Step 5projectRequired720 minutes
Image classification project
Train and evaluate a CNN image classifier on a public dataset.
Step 6readingRequired480 minutes
Object detection
Study region-based and single-stage object detection methods.
Step 7readingRequired480 minutes
Image segmentation
Study semantic, instance and panoptic segmentation.
Step 8readingRequired480 minutes
Vision Transformers
Learn patch embeddings, self-attention and modern transformer-based vision architectures.
Step 9projectRequired900 minutes
Research reproduction
Choose one computer vision paper and reproduce a key experiment or result.
Step 10milestoneRequired120 minutes
Path completion
Review your notes, projects and remaining research questions.