Generative adversarial networks for ultrasound image synthesis and analysis in nondestructive evaluation

doctoral thesis

• Title (croatian): Generativne suparničke mreže za sintezu i analizu ultrazvučnih slika u nerazornim ispitivanjima
• Author: Luka Posilović
• Mentor: Sven Lončarić (mentor)
• Committee member: Marko Subašić (predsjednik povjerenstva)
• Committee member: Zvonko Kostanjčar (član povjerenstva)
• Committee member: Zoran Kalafatić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; (Department of Electronic Systems and Information Processing); Zagreb
• Defense date and country: 2022-06-10, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing; Data Processing
• Universal decimal classification (UDC): 004 - Computer science and technology. Computing. Data processing

Abstract

Non-destructive ultrasound evaluation is a technique for flaw detection in materials. Among other things, it is used for monitoring key components of nuclear power plants, railways, and pipelines. Analysis of ultrasound data is performed by human inspectors manually analyzing acquired images. Such a process can be tedious and is highly dependent on the inspector's previous experience. Deep learning methods are hard to develop because of the lack of available data. The shortage also impacts... More training new human experts in this field, since they learn through experience. Data from real inspections can not be used because of non-disclosure agreements. On the other hand, blocks with synthetic flaws are expensive to produce. The goal of this work is to develop methods for defect detection. To improve the method’s performance, deep learning methods for generating additional synthetic images need to be developed. Generated data should be realistic and of high quality even to human experts in ultrasound evaluation. Additional data should be used to improve the performance of the deep learning defect detector. Less

Abstract (croatian)

Nedestruktivno ultrazvučno ispitivanje je metoda za detekciju pukotina u materijalima. Koristi se za praćenje kritičnih komponenata nuklearnih elektrana, željezničkih pruga, i cjevovoda. Analizu ultrazvučnih podataka najčešće ručno provode inspektori. Taj proces može biti zamoran i kvaliteta cijele inspekcije jako ovisi o prijašnjem iskustvu inspektora. Metode dubokog učenja je teško razviti zbog nedostatka podataka. Taj nedostatak također utječe na uvježbavanje novih inspektora u... More području, jer i oni uče iz iskustva. Podatci sa stvarnih inspekcija se ne mogu koristiti zbog raznih ugovora o povjerljivosti podataka. U drugu ruku, blokovi sa sintetski proizvedenim pukotinama je teško i skupo za proizvesti. Cilj ovog rada je razvoj metode za detekciju pukotina. Kako bi poboljšali rad metoda temeljenih na dubokom učenju, potrebno je razviti nove metode za generiranje dodatnih sintetskih slika. Generirani podatci moraju biti realistički i visoke kvalitete. čak i ekspertima za ultrazvučno ispitivanje. Dodatni podatci se također mogu koristiti za uneprijeđenje rada detektora pukotina temeljenog na dubokom učenju. Less

• Keywords: non-destructive testing; ultrasound image analysis; automated flaw detection; image augmentation; image generation; deep learning; generative networks
• Keywords (croatian): nerazorno ispitivanje; analiza ultrazvučnih slika; automatska detekcija defekata; augmentacija slike; generiranje slika; duboko učenje; generativne mreže
• Language: english
• URN:NBN: urn:nbn:hr:168:751803
• Promotion: 2022
• Study programme: Title: Computer Science; Study programme type: university; Study level: postgraduate; Academic / professional title: Doktor znanosti (Doktor znanosti)
• Type of resource: Text
• Extent: ix, 93 str. : graf. prikazi ; 30 cm.
• File origin: Born digital
• Access conditions: Open access
• Created on: 2022-07-07 12:20:03