Salvato in:
Dettagli Bibliografici
Autori principali: Buyukcakir, Barkin, De Tobel, Jannick, Thevissen, Patrick, Vandermeulen, Dirk, Claes, Peter
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2509.09911
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908535013507072
author Buyukcakir, Barkin
De Tobel, Jannick
Thevissen, Patrick
Vandermeulen, Dirk
Claes, Peter
author_facet Buyukcakir, Barkin
De Tobel, Jannick
Thevissen, Patrick
Vandermeulen, Dirk
Claes, Peter
contents The practical adoption of deep learning in high-stakes forensic applications, such as dental age estimation, is often limited by the 'black box' nature of the models. This study introduces a framework designed to enhance both performance and transparency in this context. We use a notable performance disparity in the automated staging of mandibular second (tooth 37) and third (tooth 38) molars as a case study. The proposed framework, which combines a convolutional autoencoder (AE) with a Vision Transformer (ViT), improves classification accuracy for both teeth over a baseline ViT, increasing from 0.712 to 0.815 for tooth 37 and from 0.462 to 0.543 for tooth 38. Beyond improving performance, the framework provides multi-faceted diagnostic insights. Analysis of the AE's latent space metrics and image reconstructions indicates that the remaining performance gap is data-centric, suggesting high intra-class morphological variability in the tooth 38 dataset is a primary limiting factor. This work highlights the insufficiency of relying on a single mode of interpretability, such as attention maps, which can appear anatomically plausible yet fail to identify underlying data issues. By offering a methodology that both enhances accuracy and provides evidence for why a model may be uncertain, this framework serves as a more robust tool to support expert decision-making in forensic age estimation.
format Preprint
id arxiv_https___arxiv_org_abs_2509_09911
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle An Autoencoder and Vision Transformer-based Interpretability Analysis of the Differences in Automated Staging of Second and Third Molars
Buyukcakir, Barkin
De Tobel, Jannick
Thevissen, Patrick
Vandermeulen, Dirk
Claes, Peter
Computer Vision and Pattern Recognition
Artificial Intelligence
68T07 (Primary)
The practical adoption of deep learning in high-stakes forensic applications, such as dental age estimation, is often limited by the 'black box' nature of the models. This study introduces a framework designed to enhance both performance and transparency in this context. We use a notable performance disparity in the automated staging of mandibular second (tooth 37) and third (tooth 38) molars as a case study. The proposed framework, which combines a convolutional autoencoder (AE) with a Vision Transformer (ViT), improves classification accuracy for both teeth over a baseline ViT, increasing from 0.712 to 0.815 for tooth 37 and from 0.462 to 0.543 for tooth 38. Beyond improving performance, the framework provides multi-faceted diagnostic insights. Analysis of the AE's latent space metrics and image reconstructions indicates that the remaining performance gap is data-centric, suggesting high intra-class morphological variability in the tooth 38 dataset is a primary limiting factor. This work highlights the insufficiency of relying on a single mode of interpretability, such as attention maps, which can appear anatomically plausible yet fail to identify underlying data issues. By offering a methodology that both enhances accuracy and provides evidence for why a model may be uncertain, this framework serves as a more robust tool to support expert decision-making in forensic age estimation.
title An Autoencoder and Vision Transformer-based Interpretability Analysis of the Differences in Automated Staging of Second and Third Molars
topic Computer Vision and Pattern Recognition
Artificial Intelligence
68T07 (Primary)
url https://arxiv.org/abs/2509.09911