TY - GEN
T1 - CheXRelNet
T2 - 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022
AU - Karwande, Gaurang
AU - Mbakwe, Amarachi B.
AU - Wu, Joy T.
AU - Celi, Leo A.
AU - Moradi, Mehdi
AU - Lourentzou, Ismini
N1 - Publisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2022
Y1 - 2022
N2 - Despite the progress in utilizing deep learning to automate chest radiograph interpretation and disease diagnosis tasks, change between sequential Chest X-rays (CXRs) has received limited attention. Monitoring the progression of pathologies that are visualized through chest imaging poses several challenges in anatomical motion estimation and image registration, i.e., spatially aligning the two images and modeling temporal dynamics in change detection. In this work, we propose CheXRelNet, a neural model that can track longitudinal pathology change relations between two CXRs. CheXRelNet incorporates local and global visual features, utilizes inter-image and intra-image anatomical information, and learns dependencies between anatomical region attributes, to accurately predict disease change for a pair of CXRs. Experimental results on the Chest ImaGenome dataset show increased downstream performance compared to baselines. Code is available at https://github.com/PLAN-Lab/ChexRelNet.
AB - Despite the progress in utilizing deep learning to automate chest radiograph interpretation and disease diagnosis tasks, change between sequential Chest X-rays (CXRs) has received limited attention. Monitoring the progression of pathologies that are visualized through chest imaging poses several challenges in anatomical motion estimation and image registration, i.e., spatially aligning the two images and modeling temporal dynamics in change detection. In this work, we propose CheXRelNet, a neural model that can track longitudinal pathology change relations between two CXRs. CheXRelNet incorporates local and global visual features, utilizes inter-image and intra-image anatomical information, and learns dependencies between anatomical region attributes, to accurately predict disease change for a pair of CXRs. Experimental results on the Chest ImaGenome dataset show increased downstream performance compared to baselines. Code is available at https://github.com/PLAN-Lab/ChexRelNet.
KW - Chest X-Ray comparison relations
KW - CXR graph representations
KW - Graph attention networks
KW - Longitudinal CXR relationships
UR - http://www.scopus.com/inward/record.url?scp=85138796677&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85138796677&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-16431-6_55
DO - 10.1007/978-3-031-16431-6_55
M3 - Conference contribution
AN - SCOPUS:85138796677
SN - 9783031164309
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 581
EP - 591
BT - Medical Image Computing and Computer Assisted Intervention – MICCAI 2022 - 25th International Conference, Proceedings
A2 - Wang, Linwei
A2 - Dou, Qi
A2 - Fletcher, P. Thomas
A2 - Speidel, Stefanie
A2 - Li, Shuo
PB - Springer
Y2 - 18 September 2022 through 22 September 2022
ER -