TY - GEN
T1 - Position-Aware Masked Autoencoder for Histopathology WSI Representation Learning
AU - Wu, Kun
AU - Zheng, Yushan
AU - Shi, Jun
AU - Xie, Fengying
AU - Jiang, Zhiguo
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.
PY - 2023
Y1 - 2023
N2 - Transformer-based multiple instance learning (MIL) framework has been proven advanced for whole slide image (WSI) analysis. However, existing spatial embedding strategies in Transformer can only represent fixed structural information, which are hard to tackle the scale-varying and isotropic characteristics of WSIs. Moreover, the current MIL cannot take advantage of a large number of unlabeled WSIs for training. In this paper, we propose a novel self-supervised whole slide image representation learning framework named position-aware masked autoencoder (PAMA), which can make full use of abundant unlabeled WSIs to improve the discrimination of slide features. Moreover, we propose a position-aware cross-attention (PACA) module with a kernel reorientation (KRO) strategy, which makes PAMA able to maintain spatial integrity and semantic enrichment during the training. We evaluated the proposed method on a public TCGA-Lung dataset with 3,064 WSIs and an in-house Endometrial dataset with 3,654 WSIs, and compared it with 6 state-of-the-art methods. The results of experiments show our PAMA is superior to SOTA MIL methods and SSL methods. The code will be available at https://github.com/WkEEn/PAMA.
AB - Transformer-based multiple instance learning (MIL) framework has been proven advanced for whole slide image (WSI) analysis. However, existing spatial embedding strategies in Transformer can only represent fixed structural information, which are hard to tackle the scale-varying and isotropic characteristics of WSIs. Moreover, the current MIL cannot take advantage of a large number of unlabeled WSIs for training. In this paper, we propose a novel self-supervised whole slide image representation learning framework named position-aware masked autoencoder (PAMA), which can make full use of abundant unlabeled WSIs to improve the discrimination of slide features. Moreover, we propose a position-aware cross-attention (PACA) module with a kernel reorientation (KRO) strategy, which makes PAMA able to maintain spatial integrity and semantic enrichment during the training. We evaluated the proposed method on a public TCGA-Lung dataset with 3,064 WSIs and an in-house Endometrial dataset with 3,654 WSIs, and compared it with 6 state-of-the-art methods. The results of experiments show our PAMA is superior to SOTA MIL methods and SSL methods. The code will be available at https://github.com/WkEEn/PAMA.
KW - Self-supervised learning
KW - WSI representation learning
UR - https://www.scopus.com/pages/publications/85174694071
U2 - 10.1007/978-3-031-43987-2_69
DO - 10.1007/978-3-031-43987-2_69
M3 - 会议稿件
AN - SCOPUS:85174694071
SN - 9783031439865
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 714
EP - 724
BT - Medical Image Computing and Computer Assisted Intervention – MICCAI 2023 - 26th International Conference, Proceedings
A2 - Greenspan, Hayit
A2 - Greenspan, Hayit
A2 - Madabhushi, Anant
A2 - Mousavi, Parvin
A2 - Salcudean, Septimiu
A2 - Duncan, James
A2 - Syeda-Mahmood, Tanveer
A2 - Taylor, Russell
PB - Springer Science and Business Media Deutschland GmbH
T2 - 26th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2023
Y2 - 8 October 2023 through 12 October 2023
ER -