Despite extensive studies on retinal layer thickness, a standardized and automated method for classifying ocular pathologies based on this data is still lacking. Many eye diseases manifest in distinct alterations in retinal microstructures, and the ability to accurately capture and interpret these changes is essential for early diagnosis and effective treatment. This thesis presents the development of an innovative framework leveraging Functional Data Analysis (FDA) to classify different eye diseases and to advance retinal health monitoring through Optical Coherence Tomography (OCT) imaging. OCT, a non-invasive and high-resolution imaging modality, enables the acquisition of detailed cross-sectional scans of the retina, cornea, and optic nerve head, making it indispensable in diagnosing and managing corneal, retinal, and optic nerve pathologies. The automated analysis of retinal layers thickness extracted from OCT images serves as a key biomarker of ocular health. Unlike previous approaches, by interpreting these thickness profiles as functional data, this project introduces a statistical modeling procedure designed to classify ocular diseases and predict health outcomes. The proposed framework offers a powerful tool for early diagnosis and personalized monitoring of retinal health. This contribution also opens up new possibilities for non-invasive, quantitative screening tools in ophthalmology.
Nonostante gli studi approfonditi sullo spessore degli strati retinici, manca ancora un metodo standardizzato e automatizzato per classificare le patologie oculari basato su questi dati. Molte malattie dell’occhio si manifestano con alterazioni specifiche nelle microstrutture retiniche, e la capacità di catturare e interpretare accuratamente questi cambiamenti è essenziale per una diagnosi precoce e un trattamento efficace. Questa tesi presenta lo sviluppo di un innovativo framework che sfrutta l’Analisi dei Dati Funzionali (FDA) per classificare diverse malattie oculari e promuovere il monitoraggio della salute retinica attraverso l’imaging con Tomografia Ottica a Coerenza (OCT). L’OCT, una modalità di imaging non invasiva e ad alta risoluzione, consente di acquisire scansioni trasversali dettagliate della retina, della cornea e della testa del nervo ottico, risultando indispensabile per la diagnosi e la gestione delle patologie corneali, retiniche e del nervo ottico. L’analisi automatizzata dello spessore degli strati retinici estratta dalle immagini OCT funge da importante biomarcatore della salute oculare. Diversamente dagli approcci precedenti, interpretando questi profili di spessore come dati funzionali, questo progetto introduce una procedura di modellizzazione statistica progettata per classificare le malattie oculari e prevedere gli esiti clinici. Il framework proposto offre uno strumento potente per la diagnosi precoce e il monitoraggio personalizzato della salute retinica. Questo contributo apre inoltre nuove prospettive per strumenti di screening quantitativi non invasivi in oftalmologia.
Functional data analysis framework for monitoring retinal health using OCT images
ESPOSITO, CHIARA
2024/2025
Abstract
Despite extensive studies on retinal layer thickness, a standardized and automated method for classifying ocular pathologies based on this data is still lacking. Many eye diseases manifest in distinct alterations in retinal microstructures, and the ability to accurately capture and interpret these changes is essential for early diagnosis and effective treatment. This thesis presents the development of an innovative framework leveraging Functional Data Analysis (FDA) to classify different eye diseases and to advance retinal health monitoring through Optical Coherence Tomography (OCT) imaging. OCT, a non-invasive and high-resolution imaging modality, enables the acquisition of detailed cross-sectional scans of the retina, cornea, and optic nerve head, making it indispensable in diagnosing and managing corneal, retinal, and optic nerve pathologies. The automated analysis of retinal layers thickness extracted from OCT images serves as a key biomarker of ocular health. Unlike previous approaches, by interpreting these thickness profiles as functional data, this project introduces a statistical modeling procedure designed to classify ocular diseases and predict health outcomes. The proposed framework offers a powerful tool for early diagnosis and personalized monitoring of retinal health. This contribution also opens up new possibilities for non-invasive, quantitative screening tools in ophthalmology.| File | Dimensione | Formato | |
|---|---|---|---|
|
2025_10_Esposito_Executive Summary.pdf
accessibile in internet per tutti
Descrizione: Executive summary
Dimensione
2.53 MB
Formato
Adobe PDF
|
2.53 MB | Adobe PDF | Visualizza/Apri |
|
2025_10_Esposito_Tesi.pdf
accessibile in internet per tutti
Descrizione: Tesi
Dimensione
12 MB
Formato
Adobe PDF
|
12 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/243744