Stable Diffusion Model-Based Scintigraphy Image Synthesis: Data Augmentation Toward Enhanced Multiclass Thyroid Diagnosis

The objective of this study is to assess the efficacy of advanced augmentation techniques, such as stable diffusion, in improving the performance of deep learning models in the classification of scintigraphic thyroid images. In this retrospective study, 2983 anterior view scintigraphic images were collected and subsequently categorized into four thyroid conditions. Both stable diffusion and conventional augmentation techniques were utilized. The generated images, alongside real images, were used to train a ResNet101V2 architecture under six different training strategies. The strategies were assessed against external datasets to evaluate model performance in terms of accuracy, precision, recall, and F1-score. The use of synthetic data in training led to consistently superior performances compared to training with only real data. Specifically, the models trained with synthetic data augmentation demonstrated higher precision and recall. The incorporation of synthetic images generated via stable diffusion significantly enhanced the diagnostic capabilities of AI models in thyroid scintigraphy interpretation. This approach not only improves the classification accuracy but also provides a viable solution to the challenge of data scarcity in medical imaging.

Authors

Institutions

Publication Details

Published
2024-09-08
DOI
https://doi.org/10.1109/euvip61797.2024.10772863
Citations
1
Primary Topic
Medical Imaging and Analysis
Type
article
Field-Weighted Citation Impact
0.18

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Stable Diffusion Model-Based Scintigraphy Image Synthesis: Data Augmentation Toward Enhanced Multiclass Thyroid Diagnosis

Yazdan Salimi, Isaac Shiri, Habib Zaidi, Ali Rasouli et al.
1 citations
Medical Imaging and Analysis
0.18
article

Stable Diffusion Model-Based Scintigraphy Image Synthesis: Data Augmentation Toward Enhanced Multiclass Thyroid Diagnosis

Yazdan Salimi, Isaac Shiri, Habib Zaidi, Ali Rasouli, Abdollah Saberi, Arman Rahmin, Seyed Rasoul Zakavi, Emran Askari, Mostafa Haghir Chehreghani, Ghasem Hajianfar, Mohsen Arabi, Atena Aghaee, Kourosh Fattahi, Soroush Bagheri, Maziar Sabouri, Ehsan Bayat, Amirhossein Sanaat, Mehdi Mogharrabi, Azin Asadzadeh
article en
1 citations

Abstract

The objective of this study is to assess the efficacy of advanced augmentation techniques, such as stable diffusion, in improving the performance of deep learning models in the classification of scintigraphic thyroid images. In this retrospective study, 2983 anterior view scintigraphic images were collected and subsequently categorized into four thyroid conditions. Both stable diffusion and conventional augmentation techniques were utilized. The generated images, alongside real images, were used to train a ResNet101V2 architecture under six different training strategies. The strategies were assessed against external datasets to evaluate model performance in terms of accuracy, precision, recall, and F1-score. The use of synthetic data in training led to consistently superior performances compared to training with only real data. Specifically, the models trained with synthetic data augmentation demonstrated higher precision and recall. The incorporation of synthetic images generated via stable diffusion significantly enhanced the diagnostic capabilities of AI models in thyroid scintigraphy interpretation. This approach not only improves the classification accuracy but also provides a viable solution to the challenge of data scarcity in medical imaging.

University of British Columbia (CA), Amirkabir University of Technology (IR), Mashhad University of Medical Sciences (IR), Golestan University (IR), Kashan University of Medical Sciences (IR), Kermanshah University of Medical Sciences (IR), University Hospital of Geneva (CH), Jahrom University of Medical Sciences (IR), Golestan University of Medical Sciences (IR), Shaheed Rajaei Cardiovascular Medical and Research Center (IR), Alborz University of Medical Sciences (IR)
Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung
Zero hunger
Openalex Percentile: Top 49%
Medical Imaging and Analysis
0.18
DOISourceOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.