20.500.12556/RUL-177984 Avtomatizacija procesa kontrole kakovosti dela radioloških inženirjev v mamografiji Automatization of radiographers'​ quality control process in mammography Uvod: Pri izdelavi visoko kakovostnih mamografskih slik ima radiološki inženir osrednjo vlogo, saj ima kvaliteta mamografskih slik zelo velik vpliv na natančnost diagnoze. V Sloveniji se za zagotavljanje visokih standardov slikanja v presejalnem programu DORA enkrat ali dvakrat letno izvaja ocenjevanje mamografskih slik radioloških inženirjev, ki delajo v tem programu. Ocenjevanje slik je zamudno, hkrati pa je ocenjevanje dokaj subjektivno, zato lahko uporaba različnih orodij za poenostavljanje postopkov ocenjevanja skrajša celoten proces in tudi doseže večjo objektivnost. Namen: Namen doktorske disertacije je preizkusiti in razviti različne postopke za avtomatizacijo vrednotenja kakovosti mamografskih slik v programu DORA ter izdelati aplikacijo z uporabo teh metod. Metode: V raziskavi smo uporabili presečno študijo z eksperimentalno metodo. Izvedbe raziskave smo se lotili v več korakih: pridobitev zbirke slik in zajem podatkov, obdelava zbirke mamografskih slik, uporaba postopkov avtomatizacije (določanje ključnih elementov na sliki, označevanje napak na slikah, uporaba postopkov segmentacije za detekcijo dojke, razvoj in preizkus delovanja konvolucijskih nevronskih mrež za prepoznavanje napake IMF, razvoj in preizkus detekcije ključnih točk na mamografskih slikah z algoritmom YOLO, izpeljava računskih postopkov za pridobitev objektivnih mer) ter končni razvoj spletne aplikacije. Rezultati in razprava: Celoten razvoj aplikacije je potekal od septembra 2020 do aprila 2024. Uspešno smo razvili in preizkusili več računalniških postopkov – od klasičnih segmentacijskih metod do naprednih modelov globokega učenja in sistemov za detekcijo objektov. Segmentacijska metoda upragovljanja se je izkazala za dovolj učinkovito pri osnovni segmentaciji dojke, medtem ko so konvolucijske nevronske mreže in modeli YOLO dosegli spodbudne rezultate pri zaznavanju in klasifikaciji posameznih kriterijev kakovosti. Zaključek: Končni rezultat naloge je izdelava spletne aplikacije za ocenjevanje kakovosti mamografskih slik. Pri primerjavi ocen ocenjevalcev in aplikacije so bili najboljši rezultati doseženi pri kriterijih z jasno vizualno definiranimi elementi, medtem ko je bila uspešnost nižja pri bolj subjektivnih merilih in redkih napakah. Introduction: Radiographers play a central role in the acquisition of high quality mammographic images, as image quality has a significant impact on diagnostic accuracy. To ensure high imaging standards within the national breast cancer screening programme DORA, image quality assessments of mammograms obtained by radiographers working in the programme are performed once or twice a year in Slovenia. This evaluation process is time-consuming and relatively subjective. Therefore, the use of automated tools to streamline the assessment procedures can both shorten the process and improve objectivity. Purpose: The aim of this dissertation is to test and develop various methods for automating the assessment of mammographic image quality within the DORA programme and to develop a dedicated web-based application that implements these methods. Methods: A cross-sectional study was conducted using an experimental approach. The study was conducted in several phases: Collection of image datasets and data acquisition; preprocessing of mammographic images; application of automation techniques (detection of key elements in the image, annotation of errors, breast segmentation methods, development and testing of convolutional neural networks [CNNs] to detect IMF errors, development and testing of object detection methods using the YOLO algorithm to detect keypoints, computational methods to determine objective quality metrics); and finally, the development of the web-based application. Results and discussion: The application development process took place between September 2020 and April 2024. We successfully developed and tested several computational approaches – from traditional segmentation methods to advanced deep learning models and object recognition systems. Thresholding proved to be sufficiently effective for basic breast segmentation, while CNNs and YOLO-based models achieved promising results in recognising and classifying individual quality criteria. Conclusion: The main outcome of this work is the development of a web application for evaluating the quality of mammographic images. When comparing the human evaluations with the application assessments, the best performance was obtained for criteria with clearly defined visual elements, while lower accuracy was observed for more subjective criteria and rare error types. mamografija kontrola kakovosti pozicioniranje avtomatizacija aplikacija subjektivnost strojno učenje v mamografiji mammography quality control positioning automatization application subjectivity machine learning in mammography true false false Slovenski jezik Angleški jezik Doktorsko delo/naloga 2026-01-15 07:15:16 2026-01-15 07:15:35 2026-02-25 08:57:16 0000-00-00 00:00:00 2026 0 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 2026-08-25 38450 8359.pdf 8359.pdf 1 41D2166A5D52E8F020E00FF598F24F96 c9d771c9db573e578c356e4cc23a73f56b5389ac9443477cd4ae9c3801dda712 9713c783-f1d9-11f0-bf92-0050569b8976 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=225459 Medicinska fakulteta 0 0 0