A Markov random field based autonomous image segmentation : bachelor's thesis

Dimitriev, Aleksandar (author), Kristan, Matej (mentor)

Image segmentation is a widely-researched topic with many algorithms available. Our goal is to segment an image, in an unsupervised way, into several coherent parts with the help of superpixels. To achieve that, we propose an iterative segmentation algorithm. The algorithm models the image by a Markov random field, whose nodes are the superpixels, and each node has both color and texture features. The superpixels are assigned labels according to their features with the help of support vector machines and the aforementioned MRF and the number of segments is iteratively reduced. The result is a segmentation of an image into several regions with requiring any user input. The segmentation algorithm was tested on a standard evaluation database, and performs on par with state-of-the-art segmentation algorithms in F-measures. In terms of oversegmentation, our approach significantly outperforms the state of the art by greatly reducing the oversegmentation of the object of interest.

• Language: English
• Keywords: segmentation, support vector machines, SVM, Markov random field, MRF, unsupervised learning, computer science, computer and information science, diploma
• Work type: Bachelor thesis/paper
• Typology: 2.11 - Undergraduate Thesis
• Organization: FRI - Faculty of Computer and Information Science
• Publisher: [A. Dimitriev]
• Year: 2014
• Number of pages: 54 str.
• PID: 20.500.12556/RUL-29528
• COBISS.SI-ID: 1536081859
• Publication date in RUL: 19.09.2014

Licences

• License: CC BY 4.0, Creative Commons Attribution 4.0 International
• Link: http://creativecommons.org/licenses/by/4.0/
• Description: This is the standard Creative Commons license that gives others maximum freedom to do what they want with the work as long as they credit the author.

Secondary language

• Language: Slovenian
• Title: Avtonomna segmentacija slik z Markovim slučajnim poljem

Abstract

Segmentacija slik je zelo raziskovano področje, za katero so na voljo številni algoritmi. Naš cilj je segmentacija slike s pomočjo superpikslov na več skladnih delov in na nenadzorovan način. Da bi to dosegli, predlagamo iterativni segmentacijski algoritem. Algoritem predstavlja sliko kot slučajno polje Markova (MRF), katerega vozlišča so superpiksli, ki imajo barvne in teksturne atribute. Superpikslom dodelimo oznake na podlagi njihovih atributov s pomočjo metode podpornih vektorjev (SVM) in že omenjenega MRF in iterativno zmanjšujemo število segmentov. Negotovo segmentacijo po vsaki iteraciji se izboljšuje in rezultat je segmentacija slike na več semantično smiselnih delov, brez pomoči uporabnika. Algoritem je bil testiran na segmentacijsko podatkovno bazo in F ocene so podobne najsodobnejšim algoritmom. Glede fragmentacije slike naš pristop bistveno prekosi stanje tehnike z zmanjšanjem števila segmentov, iz katerih je sestavljen predmet zanimanja.

• Keywords: segmentation, support vector machines, SVM, Markov random segmentacija, metoda podpornih vektorjev, SVM, slučajno polje Markova, MRF, nenadzorovano učenje, računalništvo, računalništvo in informatika, univerzitetni študij, diplomske naloge