Overview
Visual inspection and medical imaging are two applications that aim to find anything unusual in images. In this course, you’ll train and calibrate specialized models known as anomaly detectors to identify defects. You’ll also use advanced techniques to overcome common data challenges with deep learning. AI-assisted labeling is a technique to auto-label images, saving time and money when you have tens of thousands of images. If you have too few images, you’ll generate synthetic training images using data augmentation for situations where acquiring more data is expensive or impossible.
By the end of this course, you will be able to:
• Train anomaly detection models
• Generate synthetic training images using data augmentation
• Use AI-assisted annotation to label images and video files
• Import models from 3rd party tools like PyTorch
• Describe approaches to using your model outside of MATLAB
For the duration of the course, you will have free access to MATLAB, software used by top employers worldwide. The courses draw on the applications using MATLAB, so you spend less time coding and more time applying deep learning concepts.
Syllabus
- Anomaly Detection
- Train anomaly detection models. These models do not find specific objects or classes, but instead find unusual regions in images.
- Data Augmentation
- Generate synthetic images to use for training models.
- Model-Assisted Labeling
- Save hours of manual labor by using model-assisted labeling to prepare images for object detection
- Creating Your Own Models
- Learn how to diagnose problems when training models for your applications. Also, learn the options available to share and use your model outside of MATLAB.
Taught by
Amanda Wang, Matt Rich, Megan Thompson, Mehdi Alemi and Brandon Armstrong