The morphology of the dental arches is the result of a complex balance between genetic, functional, environmental, ethnic and growth-related factors, which results in such individual variability that it is difficult to define universally valid reference parameters. Quantitative knowledge of this variability within an adult population therefore plays a significant role in numerous clinical fields: from the planning of oral and maxillofacial surgery to the design of prosthetic and orthodontic appliances, right through to forensic identification. Although the three-dimensional acquisition technologies available today allow for the creation of accurate digital models of dental structures, a static geometric model describes the morphology of a single individual and does not allow for the quantification of anatomical variability within a population, the simulation of alternative configurations, or the reconstruction of missing structures on a statistically sound basis. A Statistical Shape Model (SSM) overcomes these limitations by transforming the individual 3D representation into a statistical analysis of the morphological variability of the population under study. The aim of this thesis is to validate and utilise a three-dimensional SSM of the upper and lower dental arches, constructed from 60 three-dimensional meshes (30 female, 30 male) segmented from CT images extracted from the New Mexico Decedent Image Database (NMDID). The SSM, developed using a MATLAB pipeline, is based on Generalised Procrustes Analysis and Principal Component Analysis (PCA), and enables the compact representation of anatomical variability, the synthesis of plausible shapes, and the reconstruction of missing elements. The quality of the model was assessed according to the three standard metrics established in the literature: compactness, generalisation (via leave-one-out testing and reconstruction from a partial mesh external to the dataset) and specificity. Sexual dimorphism was also analysed, separately in form space and shape space, highlighting statistically significant dimensional differences between the sexes (males have a centroid size 4–5% larger) . The results demonstrate the methodological feasibility of the proposed approach; an SSM has been obtained with potential applications in training (generation of synthetic anatomical models), prosthetic design and forensic identification, and provides a basis for further developments on larger datasets.
La morfologia delle arcate dentali è il risultato di un complesso equilibrio tra fattori genetici, funzionali, ambientali, etnici e di crescita, che si traduce in una variabilità individuale tale da rendere difficile la definizione di parametri di riferimento universalmente validi. La conoscenza quantitativa di questa variabilità, all'interno di una popolazione adulta, riveste pertanto un ruolo rilevante in numerosi ambiti clinici: dalla pianificazione di interventi di chirurgia orale e maxillofacciale alla progettazione di dispositivi protesici e ortodontici, fino all'identificazione forense. Sebbene le tecnologie di acquisizione tridimensionale oggi disponibili consentano di ottenere modelli digitali accurati delle strutture dentali, un modello geometrico statico descrive la morfologia di un singolo individuo e non consente di quantificare la variabilità anatomica di una popolazione, di simulare configurazioni alternative né di ricostruire strutture mancanti su basi statisticamente fondate. Un Modello Statistico di Forma (SSM) supera questi limiti, trasformando la rappresentazione 3D individuale in un'analisi statistica della variabilità morfologica della popolazione in esame. La presente tesi ha come obiettivo la validazione e l’utilizzo di un SSM tridimensionale delle arcate dentarie superiore e inferiore, costruito su 60 mesh tridimensionali (30 femmine, 30 maschi) segmentate da immagini TC estratte dalla New Mexico Decedent Image Database (NMDID). L’SSM, sviluppato tramite pipeline in MATLAB, è basato su Generalized Procrustes Analysis e Principal Component Analysis (PCA), consente la rappresentazione compatta della variabilità anatomica, la sintesi di forme plausibili e la ricostruzione di elementi mancanti. La qualità del modello è stata valutata secondo le tre metriche standard consolidate in letteratura: compattezza, generalizzazione (tramite test leave-one out e ricostruzione da mesh parziale esterna al dataset) e specificità. Viene analizzato anche il dimorfismo sessuale, separatamente in form space e shape space, evidenziando differenze dimensionali statisticamente significative tra i sessi (i maschi presentano un centroid size superiore del 4–5%). I risultati ottenuti dimostrano la fattibilità metodologica dell'approccio proposto, è stato ottenuto un SSM con potenziali applicazioni in ambito di training (generazione di modelli anatomici sintetici), progettazione protesica e identificazione forense, e costituisce una base per sviluppi su dataset di maggiori dimensioni.
Sviluppo di uno modello statistico di forma per l’analisi e la ricostruzione di un modello dentale completo
CARPANESE, GAIA
2025/2026
Abstract
The morphology of the dental arches is the result of a complex balance between genetic, functional, environmental, ethnic and growth-related factors, which results in such individual variability that it is difficult to define universally valid reference parameters. Quantitative knowledge of this variability within an adult population therefore plays a significant role in numerous clinical fields: from the planning of oral and maxillofacial surgery to the design of prosthetic and orthodontic appliances, right through to forensic identification. Although the three-dimensional acquisition technologies available today allow for the creation of accurate digital models of dental structures, a static geometric model describes the morphology of a single individual and does not allow for the quantification of anatomical variability within a population, the simulation of alternative configurations, or the reconstruction of missing structures on a statistically sound basis. A Statistical Shape Model (SSM) overcomes these limitations by transforming the individual 3D representation into a statistical analysis of the morphological variability of the population under study. The aim of this thesis is to validate and utilise a three-dimensional SSM of the upper and lower dental arches, constructed from 60 three-dimensional meshes (30 female, 30 male) segmented from CT images extracted from the New Mexico Decedent Image Database (NMDID). The SSM, developed using a MATLAB pipeline, is based on Generalised Procrustes Analysis and Principal Component Analysis (PCA), and enables the compact representation of anatomical variability, the synthesis of plausible shapes, and the reconstruction of missing elements. The quality of the model was assessed according to the three standard metrics established in the literature: compactness, generalisation (via leave-one-out testing and reconstruction from a partial mesh external to the dataset) and specificity. Sexual dimorphism was also analysed, separately in form space and shape space, highlighting statistically significant dimensional differences between the sexes (males have a centroid size 4–5% larger) . The results demonstrate the methodological feasibility of the proposed approach; an SSM has been obtained with potential applications in training (generation of synthetic anatomical models), prosthetic design and forensic identification, and provides a basis for further developments on larger datasets.| File | Dimensione | Formato | |
|---|---|---|---|
|
Carpanese_Gaia.pdf
Accesso riservato
Dimensione
3.68 MB
Formato
Adobe PDF
|
3.68 MB | Adobe PDF |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/108010