20.500.12556/RUL-73089 AVTOMATSKA RAZGRADNJA MAGNETNORESONANČNIH SLIK MOŽGANOV NA NORMALNE IN BOLEZENSKE STRUKTURE AUTOMATED SEGMENTATION OF MAGNETIC RESONANCE BRAIN IMAGES INTO NORMAL AND PATHOLOGICAL STRUCTURES Nevrološke in možganskožilne ter duševne bolezni so med največjimi povzročitelji telesne invalidnosti in mentalne prizadetosti v družbah razvitih držav, posledično pa imajo zelo velik in zaradi demografskih trendov še naraščajoč socialno-ekonomski učinek. Zadnje raziskave kažejo, da te bolezni vplivajo tudi na fizično strukturo možganov, npr. preko izgube nevronov, ki se odraža kot atrofija zdravih možganskih struktur ali pojav brazgotin oziroma lezij v možganskih tkivih. Trenutno je tomografsko slikanje z magnetno resonanco (MR) najbolj občutljiva tehnika za prikazovanje zdravih in patoloških možganskih struktur, v zadnjem času pa tudi za kvantitativno analizo teh struktur. V doktorski disertaciji smo se osredotočili na problem avtomatskega obrisovanja oziroma razgradnje MR slik glave na zdrave in patološke strukture. Natančna, zanesljiva in računsko učinkovita razgradnja pridobiva na pomenu v današnji klinični praksi, ker je potrebna za določanje biomarkerjev. Na podlagi MR slik določeni biomarkerji so pomembno orodje za vrednotenje, spremljanje in optimiranje zdravljenja številnih nevroloških, možganskožilnih in duševnih bolezni. Razgradnjo MR slik glave lahko naredimo z ročnim obrisovanjem posameznih struktur zanimanja, a je to opravilo precej težavno, zamudno, stroškovno neučinkovito, predvsem pa subjektivno in zato nezanesljivo. V obsežnih kliničnih študijah, ki vključujejo obdelavo velikega števila MR slik, je očitna potreba po natančnih, zanesljivih in računsko učinkovitih avtomatskih postopkih razgradnje, ki bi v primerjavi z ročnim obrisovanjem v krajšem času in konsistetno obrisali strukture, kar je pomembno za kvantitativne meritve teh struktur oziroma določanje biomarkerjev. Po drugi strani je avtomatska razgradnja zelo zahtevna in občutljiva na številne vire variabilnosti v zajemu MR slik, kot so od objekta slikanja odvisna sivinska nehomogenost zaradi različnih parametrov zajema MR slik ali razlike v kvaliteti slik med različnimi MR napravami, anatomske variabilnosti struktur zanimanja v populaciji in raznolike pojavnosti patoloških struktur. Doktorska disertacija v različnih pogledih uvaja pomembne izboljšave na področju avtomatske razgradnje MR slik in tudi širše. Za razgradnjo MR slik se pogosto uporabljajo kompleksni modeli sivinskih vrednosti večsekvenčnih MR slik, ki opisujejo številne prej omenjene vire variabilnosti MR slik. Poleg strukture modela je ključnega pomena za uporabnost modela tudi postopek za določanje parametrov tega modela iz dane MR slike, ki mora biti čimbolj natančen in robusten. V doktorski disertaciji predlagamo nov postopek za oceno modelov mešanic, ki se navadno uporabljajo za opis zdravih možganskih struktur. Novi postopek je robusten na neenakomerno vzorčenje teh struktur in na prisotnost motilnih vzorcev. Postopek je še posebej primeren za namen razgradnje možganskih struktur, ker deluje zanesljivo ne glede na velikost patoloških struktur v MR slikah. V nadaljevanju predlagamo nov način modeliranja sivinskih vrednosti večsekvenčnih MR slik, ki vključuje prostorsko porazdeljene poenostavljene modele in jih neodvisno drug od drugega določi s prej omenjenim robustnim postopkom. Združevanje tako dobljenih porazdeljenih modelov je pripeljalo do kompleksnega, a natančnega modela celotne MR slike. V zadnjem delu disertacije predlagamo nov postopek za razgradnjo zdravih in patoloških možganskih struktur na podlagi porazdeljenih modelov. Sistematično in objektivno kvantitativno vrednotenje zmogljivosti novega in še nekaterih uveljavljenih postopkov razgradnje zdravih možganskih struktur in lezij v beli možganovini na MR slikah bolnikov z multiplo sklerozo je pokazalo precejšnje povečanje natančnosti in zanesljivosti razgradnje z novim postopkom. Zaradi možnosti učinkovite paralelne implementacije je postopek računsko nezahteven. Glede na visoko natančnost, zanesljivost in učinkovitost so novi postopki primerni za določanje biomarkerjev nevroloških, možganskožilnih in duševnih bolezni. Neurodegenerative and cerebrovascular diseases, and mental disorders, are the leading cause of physical and mental disability in the population of developed countries and thus have a huge and increasing socio-economic impact. Recent researches show that these diseases and disorders also affect the physical structure of human brain, e.g. through neuronal loss that can be observed as atrophy of normal brain tissues or the occurrence of scars or lesions in the brain tissues. Currently, magnetic resonance (MR) tomographic imaging is by far the most sensitive technique for visualizing normal and pathological brain structures, but is also being increasingly used for their quantification. This Thesis is focused on the problem of automated delineation or segmentation of brain MR images into normal and pathological structures. Accurate, reliable and computationally efficient segmentation methods are nowadays required in clinical routine so as to extract various neuroimaging biomarkers, an important paraclinical tool used to characterize, monitor and optimize treatment of several neurological and cerebrovascular diseases and mental disorders. Segmentation of brain MR images can be performed manually by delineating each of the structures of interest, however, this task is cumbersome, time-consuming, expensive, but most of all subjective and thus unreliable. Especially in large clinical trials that involve processing of a large number of MR images, there is a need for accurate, reliable and computationally efficient automated segmentation methods so as to deliver timely and consistent quantitative measurements or biomarkers. Nevertheless, automated segmentation is difficult and may be hampered due to numerous sources of variability in MR images that are inherent to brain MR imaging, e.g. intensity inhomogeneity that may depend on the imaged object, differences in multi-center MR image acquisition protocols, anatomical variability across population and heterogeneity of pathology manifestation across different patients. This Thesis advances state-of-the-art in the field of automated brain MR image segmentation in several major aspects. Models of multi-sequence MR intensity of normal brain structures, which capture the aforementioned MR intensity variabilities, and accurate and robust estimation of the models are a crucial part of most state-of-the-art segmentation methods. In the Thesis a novel method is proposed for estimation of a mixture model of normal brain structures that is robust to unbalanced samples with outliers. The method is particularly suited for the estimation of MR intensity models and thus for brain structure segmentation methods, while its main advantage is high tolerance to the presence and large variations of pathological structures. Secondly, a novel model of otherwise complex whole-brain multi-sequence MR intensities is proposed that employs spatial stratification of the complex model into several simplified models and their independent, robust estimation. Subsequent recombination of the obtained simplified models resulted in accurate wholebrain MR intensity model. Thirdly, a method for segmenting normal and pathological brain structures based on locally-adaptive MR intensity model of normal-appearing brain structures is proposed. The main advantage of the method is its robustness against different imaging artifacts and anatomical variations due to a successful compilation of adaptive local modeling and robust model estimation. Finally, a quantitative and comparative evaluation of the proposed and several state-of-the-art methods for segmenting normal-appearing structures and white-matter lesions in MR images of multiple sclerosis patients, having different disease severity characterized by total lesion load, revealed that herein developed methodological contributions substantially improve the performance of segmentation of both normal appearing and pathological brain structures. Based on the observed accuracy, reliability and efficiency the proposed methods seem as good tools for the extraction of neuroimaging biomarkers. true false false Slovenski jezik Angleški jezik Doktorsko delo/naloga 2015-10-13 07:30:01 2015-10-13 07:30:03 2023-06-16 08:28:58 0000-00-00 00:00:00 2015 0 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 1970-01-01 32643 11167316 Galimzianova_Alfiia_-_AVTOMATSKA_RAZGRADNJA_MAGNETNORESONANCNIH_SLIK_MOZGANOV_NA_NORMALNE_IN_BOL.pdf Galimzianova_Alfiia_-_AVTOMATSKA_RAZGRADNJA_MAGNETNORESONANCNIH_SLIK_MOZGANOV_NA_NORMALNE_IN_BOL.pdf 1 48AF421BA6DEEE6282167F64D139F8D2 9b447e3eef6a639cfef709a4354711aef6cd6820da7232a1200758a138e43d46 6231bf0c-a1ad-11eb-a523-00155dcfd717 20.500.12556/rul/43aa11b3-13a6-4918-9c03-bd02e5f6f6b7 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=73126 Fakulteta za elektrotehniko 0 0 0