Kuvanprosessointi (TIETS05)

Julkaistu

Image Processing is a computer science and CBDA course possible to take also in English via independent studying by using a textbook.

Sisältö / Contents

Machine learning methods called neural networks and their computation algorithms are introduced. In addition, basics for deep learning are presented.

Luennot / Lectures (in Finnish)

Tiistai klo 10-12 sali X 5.3.-9.4.2019

Keskiviikko klo 10-12 sali Y 6.3.-10.4.2019

Viikoharjoitukset / Weekly assignments (in Finnish)

Torstai klo 8-10 sali Z 14.3.-11.4.2019

Proseduuri / Procedure

  • Tietojenkäsittelyopin ja CBDA:n syventävät opinnot. / Advanced study course, 5 ECTS
  • 12 luentoa (24 h), Martti Juhola
  • 5 viikkoharjoitusta / weekly exercises (10 h), Jyrki Rasku
  • Additional scores: when one makes more of all weekly exercises than 30%, additional scores [0,5] are given as follows: ≥30% 0, ≥41% 1, ≥52% 2, ≥63% 3, ≥74% 4, ≥85% 5
  • Kurssin suorittaminen / To pass the course: (1) at least 30% of weekly exercises have to be made and (2) the examination with scores of 12 or more from [0,30]

Kurssimaterliaali / Material

Luentomateriaali:

Viikkoharjoitustehtävät / Weekly assignments tasks:

Data:

Huom! / Notes!

(1) Remember to enroll yourself to the course (once opened) via NettiOpsu.

(2) Matlab is needed for weekly assignments. To load Matlab you first need an activation code and obtaining this may take time. (You also need a UTA student account.) To be done via the following link:

https://intra.uta.fi/portal/en_GB/group/info-center/software?inheritRedirect=true

(3) Remember enroll yourself on the examination (either in April or May).

 

For the students using English only:

Acquire the following book and read the relevant parts of the book (informed by Martti Juhola) by yourself. Send your solutions for the weekly assignments delivered by email to Jyrki Rasku before each Thursday (March 14, 21, 28, April 4, 11) during the course.

The book: Rafael C. Gonzalez and Richard E. Woods: Digital Image Processing, third edition.