Automatic caries detection in bitewing radiographs: Part I—deep learning

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00064165%3A_____%2F23%3A10469919" target="_blank" >RIV/00064165:_____/23:10469919 - isvavai.cz</a>

Alternative codes found

RIV/68407700:21230/23:00371239 RIV/00216208:11110/23:10469919

Result on the web

<a href="https://verso.is.cuni.cz/pub/verso.fpl?fname=obd_publikace_handle&handle=uHRM_bnn~T" target="_blank" >https://verso.is.cuni.cz/pub/verso.fpl?fname=obd_publikace_handle&handle=uHRM_bnn~T</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/s00784-023-05335-1" target="_blank" >10.1007/s00784-023-05335-1</a>

Result language

angličtina

Original language name

Automatic caries detection in bitewing radiographs: Part I—deep learning

Original language description

Objective: The aim of this work was to assemble a large annotated dataset of bitewing radiographs and to use convolutional neural networks to automate the detection of dental caries in bitewing radiographs with human-level performance.Materials and Methods: A dataset of 3989 bitewing radiographs was created and 7257 carious lesions were annotated using minimal bounding boxes. The dataset was then divided into 3 parts for the training (70%), validation (15%), and testing (15%) of multiple object detection convolutional neural networks (CNN). The tested CNN architectures included YOLOv5, Faster R-CNN, RetinaNet, and EfficientDet. To further improve the detection performance, model ensembling was used, and nested predictions were removed during post-processing. The models were compared in terms of the F1 score and average precision (AP) with various thresholds of the intersection over union (IoU).Results: The twelve tested architectures had F1 scores of 0.72-0.76. Their performance was improved by ensembling which increased the F1 score to 0.79-0.80. The best-performing ensemble detected caries with the precision of 0.83, recall of 0.77, F1 = 0.80, and AP of 0.86 at IoU=0.5. Small carious lesions were predicted with slightly lower accuracy (AP 0.82) than medium or large lesions (AP 0.88). Conclusions: The trained ensemble of object detection CNNs detected caries with satisfactory accuracy and performed at least as well as experienced dentists (see companion paper, Part II). The performance on small lesions was likely limited by inconsistencies in the training dataset.Clinical Significance: Caries can be automatically detected using convolutional neural networks. However, detecting incipient carious lesions remains challenging.

Czech name

—

Czech description

—

Type

J_imp - Article in a specialist periodical, which is included in the Web of Science database

CEP classification

—

OECD FORD branch

30208 - Dentistry, oral surgery and medicine

Project

<a href="/en/project/EF16_019%2F0000765" target="_blank" >EF16_019/0000765: Research Center for Informatics</a><br>

Continuities

V - Vyzkumna aktivita podporovana z jinych verejnych zdroju

Publication year

2023

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Name of the periodical

Clinical Oral Investigations

ISSN

1432-6981

e-ISSN

1436-3771

Volume of the periodical

27

Issue of the periodical within the volume

12

Country of publishing house

DE - GERMANY

Number of pages

9

Pages from-to

7463-7471

UT code for WoS article

001102291500001

EID of the result in the Scopus database

2-s2.0-85176781185