This thesis explains how to detect anomalies using images by learning only normal behavior. The different types of anomalies and the main approaches used are described, along with datasets and evaluation metrics. Finally, two advanced methods are explored in depth, specifically memory bank-based techniques such as PatchCore and a model that uses SSMs such as MambaAD.
Il lavoro di tesi spiega come individuare anomalie tramite l'utilizzo di immagini imparando solo il comportamento normale. Vengono descritte le diverse tipologie di anomalie e i principali approcci utilizzati, insieme a dataset e metriche di valutazione. Infine, si approfondiscono due metodi avanzati, in particolare tecniche basate su memory bank come PatchCore e un modello che fa uso di SSMs come MambaAD.
Unsupervised Anomaly Detection Approaches in Computer Vision
BONATO, GIOELE
2025/2026
Abstract
This thesis explains how to detect anomalies using images by learning only normal behavior. The different types of anomalies and the main approaches used are described, along with datasets and evaluation metrics. Finally, two advanced methods are explored in depth, specifically memory bank-based techniques such as PatchCore and a model that uses SSMs such as MambaAD.| File | Dimensione | Formato | |
|---|---|---|---|
|
Gioele_Bonato.pdf
accesso aperto
Dimensione
1.68 MB
Formato
Adobe PDF
|
1.68 MB | Adobe PDF | Visualizza/Apri |
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/20.500.12608/110884