Course syllabus - Deep Learning for Industrial Imaging, 2.5 credits
Autumn semester 2019
A1N (Second cycle, has only first-cycle course/s as entry requirements).
School of Innovation, Design and Engineering
Course literature is preliminary up to 8 weeks before course start. Course literature can be valid over several semesters.
The contents of this course are very vast and you can find a lot of information on Internet and on YouTube. However, this course followed few books but are not mandatory to buy for the course. If you want to have a nice handbook for your future then you can consider to buy them.
Digital image processing
[New ed.] : Upper Saddle River, N.J. : Prentice Hall, cop. 2002 - xx, 793 s.
ISBN: 0201180758 LIBRIS-ID: 8395912
URL: Bokens hemsida
Industrial Image Processing [Elektronisk resurs] : Visual Quality Control in Manufacturing
2nd revised ed. 2013. : Berlin, Heidelberg : Springer Berlin Heidelberg, 2013 - XVII, 369 p. 252 illus., 20 illus. in color.
ISBN: 9783642339059 LIBRIS-ID: 14809480
Cambridge, MA : MIT Press,  - xxii, 775 pages
ISBN: 978-0-262-03561-3 LIBRIS-ID: 19973915
The aim of this course is to provide students with the fundamentals of image processing and deep learning models. The student will learn to design intelligent systems using deep learning, e.g., convolutional neural network for classification, annotation, and object recognition.
After completed the course, the student will be able to
1. Demonstrate the fundamental theory of image processing.
2. Describe the fundamental needs, challenges and limitations of Big data with industrial imaging.
3. Describe and understand the basic principles of convolution neural network.
4. Demonstrate the ability to use tools for deep learning in industrial imaging
Image processing: Introduction of industrial imaging through big data and fundamentals of image processing techniques.
Deep learning with convolutional neural network: overview of neural networks as classifiers, introduction to convolutional neural networks and Deep learning architecture.
Deep learning tools: implementation of Deep learning for Image classification and object recognition, e.g. using Keras.
90 credits of which at least 60 credits in Computer Science or equivalent, including at least 15 credits in programming. In addition, Swedish course B/Swedish course 3 and English course A/English course 6 are required. For courses given entirely in English exemption is made from the requirement in Swedish course B/Swedish course 3.