Two-part lecture to introduct postgraduate students to computer vision approaches using Python and TensorFlow.
- What is a computer vision problem and how do we use machine learning to tackle them?
- What are convolutional neural networks?
- Tutorial: Building a custom dataset and learning to recognise hand-written digits
- How do we combine CNN layers into computer vision architectures?
- What do we do if we don't have as large a dataset as we'd like? Transfer learning!
- Other computer vision problems: fully-convolutional neural networks
- Tutorial: Flower species classification using Tensorflow and transfer learning