Segmentacija slika mnoštva ljudi uporabom Haralickovih teksturnih značajki

undergraduate thesis

• Title (english): Segmentation of Images of Crowd Scenes Based on Haralicks Texture Features
• Author: Filip Hustić
• Mentor: Slobodan Ribarić (mentor)
• Committee member: Slobodan Ribarić (predsjednik povjerenstva)
• Committee member: Darijan Marčetić (član povjerenstva)
• Committee member: Franjo Matković (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2019-07-12, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Tema ovog rada je segmentacija mnoštva na slikama uporabom Haralickovih teksturalnih značajki. Segmentacija se provodi promatrajući teksturalne značajke mnoštva izračunom matrica prostornih međuovisnosti sivih tonova te Haralickovih opisnika iz izračunatih matrica za dijelove slika. Izračunati opisnici predstavljaju se u obliku n-dimenzionalnog vektora opisnika koji se predaje q-NN klasifikatoru na klasifikaciju. q-NN klasifikatoru su također predane predefinirane baze vektora pomoću kojih određuje klasu vektora koje se promatra. Rezultati detekcije su prikazani kroz niz slika, a točnost detekcije je ispitana uporabom metode intersection of union.

Abstract (english)

This thesis works on segmentation of Images of Crowd Scenes Based on Haralicks Texture Features. Segmentation is done by extracting Haralicks textural features from co-occurrence matrices constructed in different parts of the picture. Extracted features are represented as a n - dimensional vector of features. Such vector is given to q-NN classifier which determines does vector describes part of picture with crowd in it or not. q-NN classifier is given a predefined base of vectors which are used in determining q-nearest neighbours of vector which will be classified. Results of detection are shown in multiple pictures. Accuracy of detection is measured using the intersection of union method.

• Keywords: mnoštvo; segmentacija; matrica prostornih međuovisnosti sivih razina; Haralickovi opisnici; q-NN klasifikator
• Keywords (english): crowd; segmentation; co-occurrence matrix; Haralick features; q-NN classifier
• Language: croatian
• URN:NBN: urn:nbn:hr:168:397906
• Study programme: Title: Computing; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2020-01-07 18:29:24