Interactive class notebooks for ECE4076 Computer Vision.
ECE4076 is a computer vision unit at Monash University, covering both classical computer vision and modern deep learning methods. These notebooks are in class activities used alongside pre-recorded lectures covering more detailed material, and paired with laboratory sessions where students implement related concepts.
- Week 1: Image handling and basic manipulation, high dimensional signals
- Week 2: Image filtering, Difference of Gaussians, Keypoint Detection, Patch Matching using SSD
- Week 3: Invariances and image transformations, assymetric feature matching
- Week 4: Camera models and homography estimation using RANSAC
- Week 5: Camera projection and vanishing points and lines
- Week 6: Multiple view geometry, space carving
- Week 7: Clustering and gaussian mixture models
- Week 8: Logistic regression
- Week 9: Understanding gradient descent
- Week 10: Object detection with Alexnet and Resnets (to be improved)
- Week 11: Object recognition with RCNNs and YOLO (to be improved)
- Week 12: Image segmentation with U-Nets (todo)
There is some additional material we unfortunately don't have time to cover in class in the extra content folder
- Vision transformers
- Generative Adversarial Networks (todo)
- Variational Autoencoders (todo)
- Clip (to be improved)
- Diffusion models (todo)