Version 1
: Received: 29 October 2024 / Approved: 29 October 2024 / Online: 30 October 2024 (11:35:19 CET)
How to cite:
Pedersen, S.; Jain, S.; Chavez, M.; Ladehoff, V.; de Freitas, B. N.; Pauwels, R. Pano-GAN: A Generative Model for Panoramic Dental Radiographs. Preprints2024, 2024102374. https://doi.org/10.20944/preprints202410.2374.v1
Pedersen, S.; Jain, S.; Chavez, M.; Ladehoff, V.; de Freitas, B. N.; Pauwels, R. Pano-GAN: A Generative Model for Panoramic Dental Radiographs. Preprints 2024, 2024102374. https://doi.org/10.20944/preprints202410.2374.v1
Pedersen, S.; Jain, S.; Chavez, M.; Ladehoff, V.; de Freitas, B. N.; Pauwels, R. Pano-GAN: A Generative Model for Panoramic Dental Radiographs. Preprints2024, 2024102374. https://doi.org/10.20944/preprints202410.2374.v1
APA Style
Pedersen, S., Jain, S., Chavez, M., Ladehoff, V., de Freitas, B. N., & Pauwels, R. (2024). Pano-GAN: A Generative Model for Panoramic Dental Radiographs. Preprints. https://doi.org/10.20944/preprints202410.2374.v1
Chicago/Turabian Style
Pedersen, S., Bruna Neves de Freitas and Ruben Pauwels. 2024 "Pano-GAN: A Generative Model for Panoramic Dental Radiographs" Preprints. https://doi.org/10.20944/preprints202410.2374.v1
Abstract
This paper presents the development of a Generative Adversarial Network (GAN) for the generation of synthetic dental panoramic radiographs to address the scarcity of data in dental research and education. A Deep Convolutional GAN (DCGAN) with Wasserstein Loss and Gradient Penalty (WGAN-GP) was trained on a dataset of 2322 radiographs of varying quality. The focus for this study was on the dentoalveolar part of the radiographs; other structures were cropped out. Significant data cleaning and preprocessing was conducted to standardize input formats while maintaining anatomical variability. Four candidate models were identified by varying the critic iterations, number of features and the use of denoising prior to training. To assess the quality of the generated images, a clinical expert evaluated generated synthetic radiographs using a ranking system based on visibility and realism from 1 (Very Poor) to 5 (Excellent). It was found that most generated radiographs showed moderate depictions of dentoalveolar anatomical structures, although they were considerably impaired by artifacts. After averaging evaluation scores, a trade-off was observed between the model trained on non-denoised data, which showed the highest subjective quality for finer structures such as the Mandibular Canal and Trabecular Bone, and one of the models trained on denoised data, which offered better overall image quality, especially for Clarity & Sharpness, and Overall Realism. The outcome serves as a foundation for further research into GAN architectures for dental applications. A GitHub repository containing the WGAN-GP model, source code, and synthetic data has been made publicly available at https://github.com/ViktorLaden/DataProjectGAN.
Medicine and Pharmacology, Dentistry and Oral Surgery
Copyright:
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.