Embedded Vision Design and Machine Learning (full-time)
Are you interested in how image processing algorithms can determine what is shown in a picture? And how to use that knowledge to create technical solutions? Join the minor in Embedded Vision and Machine Learning at HAN University of Applied Sciences!
- How do you identify the outcome of a role of the dice with a camera?
- How can you use a camera to convert hand movements into mouse control?
- How do you identify licence plate details from a moving vehicle?
- How do you apply vision algorithms to embedded systems?
- What is artificial intelligence and how do you apply this to image processing?
- What is the difference between machine learning and deep learning?
You get to delve into topics like image recognition, image processing and object classification with the help of artificial intelligence. The minor has a practical approach with different hardware platforms such as PC and embedded system. You learn to apply vision algorithms and classification methods to solve problems, both in assignments and for ideas you’ve come up with yourself.
We use software tools like OpenCV, Matlab, Qt and an IDE for developing applications for Cortex-M microcontrollers. These software tools are programmed in the programming languages Python and C/C++. You work individually and in a project team with these tools and programming languages. Your aim? To develop and produce a system that can read and process camera images and classify the found objects with the help of artificial intelligence.
Block exchange course
This is a block minor. It is offered once a year as a block in the 1st semester.
Type of exchange course
This is a specialisation minor. It allows you to specialise further within your own professional profile.
The minor consists of three subjects and a project. The three subjects are each partly theory and partly assignments. The learning outcomes are described per subject:
· Can carry out and provide reasons for the following image acquisition processes on a PC/laptop: camera, lens, lighting and interfaces.
· Can apply the following enhancement operators in OpenCV: Image algebra, geometric operators, synthetic images and contrast manipulation.
· Can apply the following segmentation operators in OpenCV: thresholding, labelling and blob measurement.
· Can apply the following feature extraction operators in OpenCV: filters, edge detection, binary morphology and colour image processing.
· Can apply the following classification operators in OpenCV: blob analysis, neural networks, blob matchers.
· The student can implement a set of vision operators in ANSI-C programming language based on a functional description and considering limitations such as performance and memory use.
· The student can solve a vision problem with an embedded system using these self-implemented operators: the classification of the three figures - circle, square and triangle. The student demonstrates the solution.
· The student can describe in a short and concise PowerPoint presentation the functional working and technical realisation of a unique operator.
· The student can create a unique vision operator in ANSI-C programming language and demonstrate how it works.
· Knows what digital image processing is, knows the backgrounds, frequency bands and is aware of the fundamental steps for digital image processing.
· Is aware of digital image fundamentals relating to the human eye, light and the electromagnetic spectrum. Is familiar with various sensors for the purpose of image acquisition and a simple image formation model.
· Is familiar with physical characteristics of optical instruments and of light and can apply these.
· Can apply the concepts of image sampling and quantization. Knows the basic relationships between pixels. Can apply mathematical tools that are important for digital image processing.
· Is familiar with intensity transformations operators (e.g. image negatives, log transformations power_law (Gamma) transformations and piecewise-linear transformations) and can apply these.
· Is familiar with histogram processing (e.g. histogram equalization, histogram matching (specification), local histogram processing and using histogram statistics for image enhancement) and can apply these.
· Is familiar with the fundamentals of spatial filtering and different filters (e.g. smoothing spatial filters and sharpening spatial filters) and can apply these.
· Is familiar with morphological image processing operators (e.g. erosion, dilation, opening, closing) and can apply these.
· Is familiar with Image segmentation operators (e.g. point, line, and edge detection) and various methods to threshold (e.g. basic global thresholding, optimum global thresholding using Otsu’s method) and can apply these.
· Can describe the shape and boundaries of a segment using statistical moments.
· Knows the application of machine learning algorithms, their training, fine-tuning and performance analysis.
· Can use tools for designing, implementing and evaluating machine learning.
· Can prepare data, train an algorithm and evaluate the performances of machine learning applied to image processing, in particular object classification.
· Knows the application of deep learning algorithms, their training, fine-tuning and performance analysis.
· Can use tools for designing, implementing and evaluating deep learning.
· Can prepare data, train an algorithm and evaluate the performances of deep learning applied to image processing, in particular object classification.
In this minor you work on all Bachelor of Engineering competences at level 3:
C8 Professional development
You work within the professional task Developing Embedded Systems.
For more information:
Tel: 06 55208761
Subscribe? Good to know!
For popular minors a draw takes place 3 to 4 weeks after the opening, if there are at that time more subscribers than available places. For the minors with places still available applies until the closing of the subscription period: Once a minor is full, it is closed!
In addition, if the number of subscribers after three weeks is below the norm; this minor may possibly be withdrawn. So if you are interested, sign up immediately!
Subscribe in time!
Note: For HAN students it means that, in case of cancellation of the first choice AFTER the period of decision (this takes about three weeks) they may re-subscribe for the still available minors.
Even then: Once a minor is full, it is closed!
A good overview of the HAN minors can be found in the minors app! The app is accessible via: http://www.minoren-han.nl/
The minor is especially suitable to students of Embedded Systems, Electrical and Electronic Engineering, Computer Technology, Information Technology or Mechatronics.
· You like working in a structured manner and think analytically.
· You have knowledge of the C/C++ programming language.
· You have knowledge and skills in one of the following subjects:
· Digital techniques such as Boolean algebra and number systems
· Data communication
· Developing software applications for PC or microcontroller
· Higher order description language such as Java, C#, Python or
related programming experience
· Working on engineering projects
Nice to know
The lessons of the minor are taught in English. If no international students register for the minor, the lessons will be taught in Dutch.
The minor is part of the Electrical and Electronic Engineering degree course. The quality of the assessment is assured by the Board of Examiners of the Institute of Engineering.
ECTS credits for this minor: 30. The assessment is divided into 12 modular exams.
· Subject 1 is assessed by means of a written exam and oral assessment.
· Subject 2 is assessed by means of two written exams.
· Subject 3 is assessed by means of two written reports.
The project is assessed by means of the delivered products, i.e. a research report, a functional design, a technical design, the realised product and two product demonstrations/presentations. The two product demonstrations/presentations count as ticks. The final grade is the average of the other eight grades for the minor.
You get two chances for each modular exam.
Classes are scheduled a maximum of four days a week. You also need to work independently on homework assignments and on the project.
- Project assignment in pairs