TY - GEN
T1 - Investigating the Impact of Image Quality on Endoscopic AI Model Performance
AU - Jaspers, Tim J. M.
AU - Boers, Tim G. W.
AU - Kusters, Carolus H. J.
AU - Jong, Martijn R.
AU - Jukema, Jelmer B.
AU - de Groof, Albert J.
AU - Bergman, Jacques J.
AU - de With, Peter H. N.
AU - van der Sommen, Fons
N1 - Publisher Copyright: © 2024, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2024
Y1 - 2024
N2 - Virtually all endoscopic AI models are developed with clean, high-quality imagery from expert centers, however, the clinical data quality is much more heterogeneous. Endoscopic image quality can degrade by e.g. poor lighting, motion blur, and image compression. This disparity between training, validation data, and real-world clinical practice can have a substantial impact on the performance of deep neural networks (DNNs), potentially resulting in clinically unreliable models. To address this issue and develop more reliable models for automated cancer detection, this study focuses on identifying the limitations of current DNNs. Specifically, we evaluate the performance of these models under clinically relevant and realistic image corruptions, as well as on a manually selected dataset that includes images with lower subjective quality. Our findings highlight the importance of understanding the impact of a decrease in image quality and the need to include robustness evaluation for DNNs used in endoscopy.
AB - Virtually all endoscopic AI models are developed with clean, high-quality imagery from expert centers, however, the clinical data quality is much more heterogeneous. Endoscopic image quality can degrade by e.g. poor lighting, motion blur, and image compression. This disparity between training, validation data, and real-world clinical practice can have a substantial impact on the performance of deep neural networks (DNNs), potentially resulting in clinically unreliable models. To address this issue and develop more reliable models for automated cancer detection, this study focuses on identifying the limitations of current DNNs. Specifically, we evaluate the performance of these models under clinically relevant and realistic image corruptions, as well as on a manually selected dataset that includes images with lower subjective quality. Our findings highlight the importance of understanding the impact of a decrease in image quality and the need to include robustness evaluation for DNNs used in endoscopy.
KW - DNN
KW - Endoscopy
KW - Image degradation
KW - Robustness
UR - http://www.scopus.com/inward/record.url?scp=85177237721&partnerID=8YFLogxK
U2 - https://doi.org/10.1007/978-3-031-47076-9_4
DO - https://doi.org/10.1007/978-3-031-47076-9_4
M3 - Conference contribution
SN - 9783031470752
VL - 14313 LNCS
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 32
EP - 41
BT - Applications of Medical Artificial Intelligence - 2nd International Workshop, AMAI 2023, Held in Conjunction with MICCAI 2023, Proceedings
A2 - Wu, Shandong
A2 - Shabestari, Behrouz
A2 - Xing, Lei
PB - Springer Science and Business Media Deutschland GmbH
T2 - 2nd International Workshop on Applications of Medical Artificial Intelligence, AMAI 2023
Y2 - 8 October 2023 through 8 October 2023
ER -