Dental implant recognition

Simple item page
dc.contributor.advisorCoetzer, Johannesen_ZA
dc.contributor.authorKohlakala, Aviween_ZA
dc.contributor.otherStellenbosch University. Faculty of Science. Dept. of Mathematical Sciences. Applied Mathematics Division.en_ZA
dc.date.accessioned2023-08-18T08:41:35Zen_ZA
dc.date.accessioned2024-01-08T11:59:39Zen_ZA
dc.date.available2023-08-18T08:41:35Zen_ZA
dc.date.available2024-01-08T11:59:39Zen_ZA
dc.date.issued2023-09en_ZA
dc.descriptionThesis (PhD)--Stellenbosch University, 2023.en_ZA
dc.description.abstractENGLISH ABSTRACT: Deep learning-based frameworks have recently been steadily outperforming existing state-of-the-art systems in a number of computer vision applications, but these models require a large number of training samples in order to effectively train the model parameters. Within the medical field the limited availability of training data is one of the main challenges faced when using deep learning to create practical clinical applications in medical imaging. In this dissertation a novel algorithm for generating artificial training samples from triangulated three-dimensional (3D) surface models within the context of dental implant recognition is proposed. The proposed algorithm is based on the calculation of two-dimensional (2D) parallel projections from a number of different angles of 3D volumetric representations of computer-aided design (CAD) surface models. A fully convolutional network (FCN) is subsequently trained on the artificially generated X-ray images for the purpose of automatically identifying the connection type associated with a specific dental implant in an actual X-ray image. An ensemble of image processing and deep learning-based techniques capable of distinguishing between pixels that belong to an implant from those belonging to the background in an actual X-ray image is developed. Normalisation and preprocessing techniques are subsequently applied to the segmented dental implants within the questioned actual X-ray image. The normalised dental implants are presented to the trained FCN for classification purposes. Experiments are conducted on two data sets that contain the simulated and actual X-ray images in order to gauge the proficiency of the proposed systems. Given the fact that the novel systems proposed in this study utilise an ensemble of techniques that has not been employed for the purpose of dental implant classification/recognition on any previous occasion, the results achieved in this study are encouraging and constitute a significant contribution to the current state of the art, especially in scenarios where the proposed systems are combined with existing systems.en_ZA
dc.description.abstractAFRIKAANS OPSOMMING: Diepleergebaseerde raamwerke het onlangs op bestaande staat-van-die-kuns stelsels in ’n aantal rekenaarvisietoepassings begin verbeter, maar hierdie modelle verg ’n groot aantal voorbeelde ten einde die modelparameters effektief af te rig. In die mediese veld is die beperkte beskikbaarheid van afrigdata een van die hoofuitdagings vir praktiese kliniese toepassings in mediese beeldvorming. In hierdie proefskrif word ’n nuwe algoritme vir die skep van kunsmatige afrigvoorbeelde vanuit driehoekgebaseerde driedimensionele (3D) oppervlakmodelle binne die konteks van tandimplantaatherkenning voorgestel. Die voorgestelde algoritme is op die berekening van tweedimensionele (2D) parallelle projeksies vanuit ’n aantal verskillende hoeke van 3D volumetriese voorstellings van rekenaarmatige ontwerp (CAD) oppervlakmodelle gebaseer. ’n Vol konvolusie-netwerk (FCN) word vervolgens op die kunsmatig-gegenereerde X-straalbeelde afgerig met die doel om die verbindingstipe geassosieer met ’n spesifieke tandimplantaat te identifiseer. ’n Ensemble van beeldverwerkings- en diepleergebaseerde tegnieke, wat in staat is om piksels wat tot ’n implantaat in ’n werklike X-straalbeeld hoort van dié wat tot die agtergrond hoort te onderskei, word ontwikkel. Normalisasie en voorverwerkingstegnieke word vervolgens op die gesegmenteerde tandimplantate in ’n bevraagtekende werklike X-straalbeeld toegepas. Die genormaliseerde tandimplantate word aan die afgerigte FCN voorgelê vir klassifikasiedoeleindes. Eksperimente word op twee datastelle, wat gesimuleerde en werklike X-straalbeelde bevat, toegepas ten einde die vaardigheid van die voorgestelde stelsels te beraam. Gegee die feit dat die nuwe stelsels voorgestel in hierdie studie van ’n ensemble van tegnieke gebruik maak wat nog nie voorheen vir die doel van tandimplantaatherkenning gebruik is nie, is die resultate behaal in hierdie studie baie bemoedigend en is dit ’n beduidende bydrae tot die huidige staat van die kuns, veral in scenarios waar die voorgestelde stelsels met bestaande stelsels gekombineer word.en_ZA
dc.description.versionDoctorateen_ZA
dc.format.extentxvii, 113 pages : illustrationsen_ZA
dc.identifier.urihttps://scholar.sun.ac.za/handle/10019.1/128811en_ZA
dc.language.isoen_ZAen_ZA
dc.language.isoen_ZAen_ZA
dc.publisherStellenbosch : Stellenbosch Universityen_ZA
dc.rights.holderStellenbosch Universityen_ZA
dc.subject.lcshDeep learning (Machine learning)en_ZA
dc.subject.lcshComputer vision -- Industrial applicationsen_ZA
dc.subject.lcshDiagnostic imagingen_ZA
dc.subject.lcshDental implantsen_ZA
dc.subject.lcshDigital elevation modelsen_ZA
dc.subject.lcshPattern perceptionen_ZA
dc.titleDental implant recognitionen_ZA
dc.typeThesisen_ZA
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
kohlakala_dental_2023.pdf
Size:
8.36 MB
Format:
Adobe Portable Document Format
Description:
Download
Collections
Doctoral Degrees (Applied Mathematics)