Learning outcomes |
The aim of the course is to give the student a cross-section of the central theory of machine vision, methods with different algorithms, and applications so that he / she can independently apply machine vision methods in solving practical problems as well as solving theoretical problems.
Student
- knows the central methods of machine vision and their application areas, as well as the limitations, understands the central differences between the latest and the classic methods
- identifies which methods are worthwhile or can be used and which are not useful for solving a particular problem
- nows the pros and cons of different methods and related algorithms
- is able to apply machine vision techniques to various problems that are increasingly encountered in different areas of society .
|
Content |
The course tackles for example the following areas:
- History of machine vision
- Digital image and its formation, different picture grids
- Binary image, grayscale image, color image, spectral image
- The effect of image algebra and grid metrics on methods
- Mathematical morphology in machine vision
- * Central methods such as dilation, erosion, opening, closing, watershed transform, etc.
- Distance transforms for both binary, grayscale, and color images
- Color spaces and their use in machine vision
- Edge detection methods
- Geometric pattern recognition (Hough transform, etc.), corner detection methods
- Spatial and frequency-domain images
- * Frequency filtering of images
- * Different types of image noise and corresponding filtering methods
- Fourier Transform in machine vision (for both grayscale and color images)
- Detection of physical objects using both spatial and frequency-domain images
- Image compression
- * Lossy and lossless compression
- * Image compression methods based on image transforms (eg JPEG and JPEG2000)
- * Image compression methods based on prediction
|
Modes of study |
Contact teaching: Lectures 32 hours, exercises 16 hours, independent work 110 hours.
Lectures and exercises are not compulsory, but participation is recommended. They go through methods that are helpful when doing the home assignment.
Requirements: Exam (grade 0-5), assignment (grade: rejected / approved / approved with distinction). Home assignment grade approved with distinction increases the exam grade to + 1, and makes that the final grade, but not from 0 to 1.
The area of the exam are lectures and exercises.
Minority students, Open University students, and Continuing Learners take the course by passing an exam on a general exam day, consisting of lectures (PP slides available) and completing the home assignment as approved |
Teaching methods |
Both teaching and exercises are given by contact teaching |
Study materials |
The lecture will cover material compiled by the lecturer from different sources. Some of the content can be found in Rafael C. Gonzalez, Richard E. Woods: Digital Image Processing, 2nd Edition, Prentice-Hall, 2002 or later. |
Evaluation criteria |
Passing the course requires passing the exam with a grade of at least 1, and completing the assignment with a grade of at least approved. The exam is graded on a scale of 0-5. The assignment will be graded on the scale of rejected / approved / approved with distinction. |
Time |
Spring semester |
Offering data |
Computer science M.Sc students, minor subject students, Computer science international M.Sc students, Computer science exchange students, students whom have done computer science basic and intermediate studies |
Further information |
Teaching in English. The language of instruction is English. A Finnish graduate student can take the exam and / or assignment in Finnish. |
|