Cross-scale multi-instance learning for pathological image diagnosis
Open accessAbstract
Analyzing high resolution whole slide images (WSIs) with regard to information across multiple scales poses a significant challenge in digital pathology. Multi-instance learning (MIL) is a common solution for working with high resolution images by classifying bags of objects (i.e. sets of smaller image patches). However, such processing is typically performed at a single scale (e.g., 20× magnification) of WSIs, disregarding the vital inter-scale information that is key to diagnoses by human pathologists. In this study, we propose a novel cross-scale MIL algorithm to explicitly aggregate inter-scale relationships into a single MIL network for pathological image diagnosis. The contribution of this paper is three-fold: (1) A novel cross-scale MIL (CS-MIL) algorithm that integrates the multi-scale information and the inter-scale relationships is proposed; (2) A toy dataset with scale-specific morphological features is created and released to examine and visualize differential cross-scale attention; (3) Superior performance on both in-house and public datasets is demonstrated by our simple cross-scale MIL strategy. The official implementation is publicly available at https://github.com/hrlblab/CS-MIL.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In integrating multi-scale histopathology images, methods like cross-scale attention mechanisms and multiple instance learning have been effective.
“In other cases, such as integrating multi-scale histopathology images, methods like cross-scale attention mechanisms and multiple instance learning (MIL) have been effective [174].”
From Challenges in AI-driven Biomedical Multimodal Data Fusion and Analysis (2025) · cites this paper as [174]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
Deep Learning using Rectified Linear Units (ReLU)Preprint
Agarap, A. F. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1803.08375 · Cited by 3,792 · cited 2× · see citing works
“Cross-scale features ( fs) are simultaneously fed into a cross-scale multi-instance learning network (CS-MIL), which contains two fully convolutional layers with kernel size of 1×1, and an ReLU activation function.”
Exploring Simple Siamese Representation Learning
Chen, X., & He, K. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.01549 · Cited by 3,725 · cited 1× · see citing works
“Our approach follows a similar design by utilizing our dataset to train a contrastive-learning model, SimSiam (Chen and He, 2021), as a phenotype encoder (Es) to extract high-level phenotype features (Fs) from patches (Is), as shown in equation 1.”
MethodEhteshami Bejnordi, B., Veta, M., Johannes van Diest, P., van Ginneken, B., Karssemeijer, N., Litjens, G., van der Laak, J. A. W. M., and the CAMELYON16 Consortium, Hermsen, M., Manson, Q. F., et al. (2017) · JAMA · doi:10.1001/jama.2017.14585 · Cited by 3,293 · see citing works
Clinical-grade computational pathology using weakly supervised deep learning on whole slide images
Campanella, G., Hanna, M. G., Geneslaw, L., Miraflor, A., Werneck Krauss Silva, V., Busam, K. J., Brogi, E., Reuter, V. E., Klimstra, D. S., & Fuchs, T. J. (2019) · Nature Medicine · doi:10.1038/s41591-019-0508-1 · Cited by 2,799 · see citing works
Data-efficient and weakly supervised computational pathology on whole-slide images
Lu, M. Y., Williamson, D. F. K., Chen, T. Y., Chen, R. J., Barbieri, M., & Mahmood, F. (2021) · Nature Biomedical Engineering · doi:10.1038/s41551-020-00682-w · Cited by 2,282 · see citing works
Deep Learning for Identifying Metastatic Breast CancerPreprint
Wang, D., Khosla, A., Gargeya, R., Irshad, H., & Beck, A. H. (2016) · arXiv.org · doi:10.48550/arxiv.1606.05718 · Cited by 1,028 · cited 1× · see citing works
“Recent MIL based approaches have greatly benefited from using deep neural networks for feature extraction and aggregation (Ilse et al., 2018; Wang et al., 2016; Oquab et al., 2015).”
BackgroundLi, B., Li, Y., & Eliceiri, K. W. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.01409 · Cited by 957 · cited 7× · see citing works
“, (Li et al., 2021) employed a feature concatenation strategy, where high-level features of each region from different scales were merged to in-”
BackgroundMethodInfluentialPatch-Based Convolutional Neural Network for Whole Slide Tissue Image Classification
Hou, L., Samaras, D., Kurç, T., Gao, Y., Davis, J., & Saltz, J. H. (2016) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2016.266 · Cited by 902 · cited 4× · see citing works
“However, annotating images pixel- or patchwise is computationally expensive for a standard supervised deep learning system (Hou et al., 2016; Mousavi et al., 2015; Maksoud et al., 2020; Dimitriou et al., 2019).”
BackgroundMethodInfluentialIs object localization for free? - Weakly-supervised learning with convolutional neural networks
Oquab, M., Bottou, L., Laptev, I., & Sivic, J. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298668 · Cited by 886 · cited 1× · see citing works
“Recent MIL based approaches have greatly benefited from using deep neural networks for feature extraction and aggregation (Ilse et al., 2018; Wang et al., 2016; Oquab et al., 2015).”
BackgroundAI-based pathology predicts origins for cancers of unknown primary
Lu, M. Y., Chen, T. Y., Williamson, D. F. K., Zhao, M., Shady, M., Lipkova, J., & Mahmood, F. (2021) · Nature · doi:10.1038/s41586-021-03512-4 · Cited by 636 · see citing works
Deep learning for prediction of colorectal cancer outcome: a discovery and validation study
Skrede, O.-J., De Raedt, S., Kleppe, A., Hveem, T. S., Liestøl, K., Maddison, J., Askautrud, H. A., Pradhan, M., Nesheim, J. A., Albregtsen, F., et al. (2020) · The Lancet · doi:10.1016/s0140-6736(19)32998-8 · Cited by 627 · see citing works
Scaling Vision Transformers to Gigapixel Images via Hierarchical Self-Supervised Learning
Chen, R. J., Chen, C., Li, Y., Chen, T. Y., Trister, A. D., Krishnan, R. G., & Mahmood, F. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.01567 · Cited by 621 · see citing works
Yao, J., Zhu, X., Jonnagaddala, J., Hawkins, N., & Huang, J. (2020) · Medical Image Analysis · doi:10.1016/j.media.2020.101789 · Cited by 522 · cited 10× · see citing works
“For example, DeepAttnMISL (Yao et al., 2020) used MIL to cluster image patches into different “bags” to model and aggregate diverse local features for patient-level diagnosis.”
MethodResultInfluentialZhang, H., Meng, Y., Zhao, Y., Qiao, Y., Yang, X., Coupland, S. E., & Zheng, Y. (2022) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52688.2022.01824 · Cited by 509 · cited 2× · see citing works
“, 2021); (8) A Double-Tier Feature Distillation when aggregating features from multiple scales and multiple locations (Zhang et al., 2022); (9) a Hierarchical Image Pyramid Transformer (HIPT) with self-supervised learning (Chen et al.”
MethodDeep Learning for Whole Slide Image Analysis: An Overview
Dimitriou, N., Arandjelović, O., & Caie, P. D. (2019) · Frontiers in Medicine · doi:10.3389/fmed.2019.00264 · Cited by 334 · see citing works
WSISA: Making Survival Prediction from Whole Slide Histopathological Images
Zhu, X.-L., Yao, J., Zhu, F.-Y., & Huang, J.-Z. (2017) · Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2017.725 · Cited by 286 · cited 1× · see citing works
“It is noted that high-level features are more comprehensive than low-resolution thumbnail images in representing phenotypes (Zhu et al., 2017).”
BackgroundHashimoto, N., Fukushima, D., Koga, R., Takagi, Y., Ko, K., Kohno, K., Nakaguro, M., Nakamura, S., Hontani, H., & Takeuchi, I. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.00391 · Cited by 251 · cited 6× · see citing works
“…al., (Hou et al., 2016) proposed a decision fusion model that aggregated patch-level predictions generated by patch-level CNNs. Hashimoto et al., (Hashimoto et al., 2020b) proposed a novel CNN-based technique for cancer subtype classification by effectively merging multiple-instance, domain…”
BackgroundMethodInfluentialYeshi, K., Ruscher, R., Hunter, L., Daly, N. L., Loukas, A. C., & Wangchuk, P. (2020) · Journal of Clinical Medicine · doi:10.3390/jcm9051273 · Cited by 190 · see citing works
Adaptive Weighting Multi-Field-Of-View CNN for Semantic Segmentation in Pathology
Tokunaga, H., Teramoto, Y., Yoshizawa, A., & Bise, R. (2019) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2019.01288 · Cited by 149 · cited 2× · see citing works
“C V ] 1 A pr 2 02 3 pyramidal) nature of WSIs, which can consist of scales from , thereby allowing pathologists to examine both local and global morphological features (Bejnordi et al., 2015; Gao et al., 2016; Tokunaga et al., 2019).”
BackgroundGubatan, J., Levitte, S., Patel, A., Balabanis, T., Wei, M. T., & Sinha, S. R. (2021) · World Journal of Gastroenterology · doi:10.3748/wjg.v27.i17.1920 · Cited by 121 · see citing works
Negative Log Likelihood Ratio Loss for Deep Neural Network Classification
Yao, H.-S., Zhu, D., Jiang, B., & Peng, Y. (2018) · Advances in Intelligent Systems and Computing · doi:10.1007/978-3-030-32520-6_22 · Cited by 104 · cited 1× · see citing works
“The Negative LogLikelihood Loss (Yao et al., 2019) was employed to compare the slide-wise prediction (Ypred) for the bag with the weakly label in 6.”
MethodAutomated discrimination of lower and higher grade gliomas based on histopathological image analysis
Mousavi, H. S., Monga, V., Rao, G., & Rao, A. (2015) · Journal of Pathology Informatics · doi:10.4103/2153-3539.153914 · Cited by 86 · cited 1× · see citing works
“However, annotating images pixel- or patchwise is computationally expensive for a standard supervised deep learning system (Hou et al., 2016; Mousavi et al., 2015; Maksoud et al., 2020; Dimitriou et al., 2019).”
BackgroundGordon, I., Bettenworth, D., Bokemeyer, A., Srivastava, A., Rosty, C., Hertogh, G., Robert, M., Valasek, M. A., Mao, R., Kurada, S., et al. (2019) · Gastroenterology · doi:10.1053/j.gastro.2019.08.033 · Cited by 78
Bejnordi, B., Litjens, G., Hermsen, M., Karssemeijer, N., & van der Laak, J. A. (2015) · Medical Imaging · doi:10.1117/12.2081768 · Cited by 54 · see citing works
Barbano, C. A., Perlo, D., Tartaglione, E., Fiandrotti, A., Bertero, L., Cassoni, P., & Grangetto, M. (2021) · Proceedings - International Conference on Image Processing · doi:10.1109/icip42928.2021.9506198 · Cited by 52 · see citing works
Self-Supervision Closes the Gap Between Weak and Strong Supervision in HistologyPreprint
Dehaene, O., Camara, A., Moindrot, O., de Lavergne, A., & Courtiol, P. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2012.03583 · Cited by 47 · see citing works
Kraszewski, S., Szczurek, W., Szymczak, J. E., Reguła, M., & Neubauer, K. (2021) · Journal of Clinical Medicine · doi:10.3390/jcm10204745 · Cited by 47 · cited 1× · see citing works
“With the rapid advancements in whole slide imaging and deep learning techniques, the potential for computer-assisted clinical diagnosis and exploration in digital pathology (Kraszewski et al., 2021; Con et al., 2021; Kiyokawa et al., 2022; Syed and Stidham, 2020) is rapidly increasing, making it a promising”
BackgroundKiyokawa, H., Abe, M., Matsui, T., Kurashige, M., Ohshima, K., Tahara, S., Nojima, S., Ogino, T., Sekido, Y., Mizushima, T., et al. (2022) · The American Journal of Pathology · doi:10.1016/j.ajpath.2022.03.006 · Cited by 32 · cited 1× · see citing works
“With the rapid advancements in whole slide imaging and deep learning techniques, the potential for computer-assisted clinical diagnosis and exploration in digital pathology (Kraszewski et al., 2021; Con et al., 2021; Kiyokawa et al., 2022; Syed and Stidham, 2020) is rapidly increasing, making it a promising”
BackgroundCon, D., van Langenberg, D. V., & Vasudevan, A. (2021) · World Journal of Gastroenterology · doi:10.3748/wjg.v27.i38.6476 · Cited by 31 · see citing works
SOS: Selective Objective Switch for Rapid Immunofluorescence Whole Slide Image Classification
Maksoud, S., Zhao, K., Hobson, P., Jennings, A., & Lovell, B. (2020) · Computer Vision and Pattern Recognition · doi:10.1109/cvpr42600.2020.00392 · Cited by 30 · cited 1× · see citing works
“However, annotating images pixel- or patchwise is computationally expensive for a standard supervised deep learning system (Hou et al., 2016; Mousavi et al., 2015; Maksoud et al., 2020; Dimitriou et al., 2019).”
BackgroundMulti-scale learning based segmentation of glands in digital colonrectal pathology images
Gao, Y., Liu, W., Arjun, S., Zhu, L.-J., Ratner, V., Kurç, T., Saltz, J., & Tannenbaum, A. (2016) · SPIE Medical Imaging · doi:10.1117/12.2216790 · Cited by 18 · see citing works
Chen, J.-N., Cheung, H., Milot, L., & Martel, A. L. (2020) · International Conference on Medical Image Computing and Computer-Assisted Intervention · doi:10.1007/978-3-030-87240-3_72 · Cited by 17
Syed, S., & Stidham, R. W. (2020) · Inflammatory Bowel Diseases · doi:10.1093/ibd/izaa211 · Cited by 16 · cited 1× · see citing works
“With the rapid advancements in whole slide imaging and deep learning techniques, the potential for computer-assisted clinical diagnosis and exploration in digital pathology (Kraszewski et al., 2021; Con et al., 2021; Kiyokawa et al., 2022; Syed and Stidham, 2020) is rapidly increasing, making it a promising”
BackgroundDeng, R.-N., Cui, C., Remedios, L. W., Bao, S.-X., Womick, R. M., Chiron, S., Li, J., Roland, J. T., Lau, K. S., Liu, Q., et al. (2022) · MMMI@MICCAI · doi:10.48550/arxiv.2208.07322 · Cited by 16
Self-citationBao, S.-X., Chiron, S., Tang, Y.-C., Heiser, C. N., Southard-Smith, A. N., Lee, H. H., Ramirez, M., Huo, Y.-K., Washington, M., Scoville, E. A., et al. (2021) · Medical Imaging · doi:10.1117/12.2581074 · Cited by 10 · see citing works
Self-citation
Flagged first, then by citations
6 of the 41 references couldn’t be matched to a known paper.
Research works printed but not linked: 6
- Barbano CA, Perlo D, Tartaglione E, Fiandrotti A, Bertero L, Cassoni P, Grangetto M, 2021a. Unitopatho, a labeled histopathological dataset for colorectal polyps classification and adenoma dysplasia grading, in: 2021 IEEE International Conference on Image Processing (ICIP), IEEE. pp. 76– 80.
- Hashimoto N, Fukushima D, Koga R, Takagi Y, Ko K, Kohno K, Nakaguro M, Nakamura S, Hontani H, Takeuchi I, 2020b. Multi-scale domain-adversarial multiple-instance cnn for cancer subtype classification with unannotated histopathological images, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 3852– 3861.
- Ilse M, Tomczak J, Welling M, 2018. Attention-based deep multiple instance learning, in: International conference on machine learning, PMLR. pp. 2127– 2136.
- Schirris Y, Gavves E, Nederlof I, Horlings HM, Teuwen J, 2021. Deepsmile: Self-supervised heterogeneity-aware multiple instance learning for dna damage response defect classification directly from h&e whole-slide images. arXiv preprint arXiv:2107.09405. · arXiv
- Wang S, Zhu Y, Yu L, Chen H, Lin H, Wan X, Fan X, Heng PA, 2019. Rmdl: Recalibrated multi-instance deep learning for whole slide gastric image classification. Medical image analysis 58, 101549. 31499320
- Xiong C, Chen H, Sung J, King I, 2023. Diagnose like a pathologist: Transformer-enabled hierarchical attention-guided multiple instance learning for whole slide image classification. arXiv preprint arXiv:2301.08125 · arXiv
Cited by
Sorted by citations
Exploring Multiple Instance Learning (MIL): A brief survey
Waqas, M., Ahmed, S. U., Tahir, M. A., Wu, J., & Qureshi, R. (2024) · Expert systems with applications · doi:10.2139/ssrn.4728166 · Cited by 87
Challenges in AI-driven Biomedical Multimodal Data Fusion and Analysis
Liu, J., Cen, X., Yi, C., Wang, F.-A., Ding, J., Cheng, J., Wu, Q., Gai, B., Zhou, Y., He, R., et al. (2025) · Genomics, Proteomics & Bioinformatics · doi:10.1093/gpbjnl/qzaf011 · Cited by 75 · see citing works
Khosravi, P., Fuchs, T. J., & Ho, D. J. (2025) · Cancer Research · doi:10.1158/0008-5472.can-24-3630 · Cited by 41 · see citing works
Large-vocabulary forensic pathological analyses via prototypical cross-modal contrastive learning
Shen, C., Lian, C.-F., Zhang, W., Wang, F., Zhang, J., Fan, S., Wei, X., Wang, G., Li, K., Mu, H., et al. (2024) · Nature Communications · doi:10.1038/s41467-025-62060-x · Cited by 17 · see citing works
“…advancements in SSL and foundation models within the broader machine learning community, recent CPath studies have begun to pre-train models using a variety of unlabeled data, subsequently fi ne-tuning them for speci fi c downstream tasks, exemplifying the SSL-based transfer learning paradigm 31,32 .”
Guan, X.-Y., Zhu, Q., Sun, L., Zhao, J.-Y., Zhang, D.-Q., Wan, P., & Shao, W. (2024) · Pattern Recognition · doi:10.1016/j.patcog.2024.110696 · Cited by 14
Dual structure-aware consensus graph learning for incomplete multi-view clustering
Sun, L., Wong, W., Fu, Y.-S., Wen, J., Li, M., Lu, Y.-W., & Fei, L. (2025) · Pattern Recognition · doi:10.1016/j.patcog.2025.111582 · Cited by 13 · see citing works
Zhao, Z.-H., Chen, D., Wang, R., Zhang, X.-K., Wen, X.-B., Zheng, X.-Y., Liu, S.-S., Chen, H., Zhang, Y., Huang, D., et al. (2026) · PLoS Medicine · doi:10.1371/journal.pmed.1004614 · Cited by 12 · see citing works
“Incorporating compensation mechanisms for low-abundance signals may further enhance the performance of WSINet [53,54].”
Ayaz, H., Sümer, E., Öztürk-Işık, E., Booth, T. C., Tormey, D., McLoughlin, I., & Unnikrishnan, S. (2025) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2025.110649 · Cited by 6 · see citing works
Jin, H., Shen, J., Cui, L., Shi, X., Li, K., & Zhu, X. (2025) · Medical Image Analysis · doi:10.1016/j.media.2025.103468 · Cited by 11 · see citing works
Xie, Y., Liu, Z., Zhao, J., & Ma, J. (2024) · Information Fusion · doi:10.1016/j.inffus.2024.102847 · Cited by 11 · see citing works
Xie, Y., Liu, Z., Chen, J., Zhang, W., Zhao, J., & Ma, J. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.126831 · Cited by 8 · see citing works
Zhang, R., Wang, K., Wang, S., Wang, C., Cao, T., Ci, C., Xu, M., & Ge, M. (2025) · Breast Cancer Research · doi:10.1186/s13058-025-02129-z · Cited by 13 · see citing works
“Inspired by the recent success of multi-instance learning (MIL) methods for histopathology, we combined MIL with a deep learning network architecture to extract deep learning features from MRI images [6–8].”
PTCMIL: Multiple Instance Learning via Prompt Token Clustering for Whole Slide Image Analysis
Zhao, B.-D., Kim, S., Chen, H., Zhou, C., Gao, Z.-H., Wang, G., & Li, X.-X. (2025) · International Conference on Medical Image Computing and Computer-Assisted Intervention · doi:10.48550/arxiv.2507.18848 · Cited by 9
Enhancing tumor deepfake detection in MRI scans using adversarial feature fusion ensembles
Ali, A. A., Basha, H. A., Thanuja, K., Puneet, P., Gupta, S. K., & Kim, S. (2025) · Scientific Reports · doi:10.1038/s41598-025-31231-7 · Cited by 8 · see citing works
Applications of artificial intelligence in digital pathology for gastric cancer
Chen, S., Ding, P., Guo, H., Meng, L., Zhao, Q., & Li, C. (2024) · Frontiers in Oncology · doi:10.3389/fonc.2024.1437252 · Cited by 12 · see citing works
Lu, X., Gong, X., & Chen, Y. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.127978 · Cited by 4 · see citing works
Koga, R., Yokota, T., Arihiro, K., & Hontani, H. (2024) · Journal of Pathology Informatics · doi:10.1016/j.jpi.2024.100413 · Cited by 6 · see citing works
Learning with less supervision: A survey of label-efficient learning for medical image analysis
Jin, C., Guo, Z., Lin, Y., Luo, L., & Chen, H. (2026) · Medical Image Analysis · doi:10.1016/j.media.2026.104062 · Cited by 31 · see citing works
Inamdar, V., & Shaila, S. (2026) · International Journal of Breast Cancer · doi:10.1155/ijbc/5948413 · Cited by 5 · see citing works
Classification of Whole-Slide Pathology Images Based on State Space Models and Graph Neural Networks
Ding, F., Cai, C., Li, J., Liu, M., Jiao, Y., Wu, Z., & Xu, J. (2025) · Electronics · doi:10.3390/electronics14102056 · Cited by 6 · see citing works
Self-citationOut-of-distribution generalization for segmentation of lymph node metastasis in breast cancer
Varnava, Y., Jakate, K., Garnett, R., Androutsos, D., Tyrrell, P., & Khademi, A. (2025) · Scientific Reports · doi:10.1038/s41598-024-80495-y · Cited by 5 · see citing works
“Therefore, incorporating multi-resolution architectures 62,73–78 could provide the necessary context for more accurate segmentations.”
Idaji, M. J., Hense, J., Neuhauser, T., Krause, A., Luo, Y., Eberle, O., Schnake, T., Ciernik, L., Jafari, F. R., Vahidimajd, R., et al. (2026) · Medical Image Analysis · doi:10.1016/j.media.2026.104148 · Cited by 1 · see citing works
“Some suggest modifications in the model architecture or training process to enhance the quality of the attention heatmaps [115, 48, 78, 125, 49] or combine attention with additional model outputs such as patch prediction scores [126, 47, 127, 18].”
Liu, F., Liang, S.-N., Jaward, M. H., Ong, H. F., & Wang, H. (2025) · Computerized Medical Imaging and Graphics · doi:10.1016/j.compmedimag.2025.102664 · Cited by 3 · see citing works
Robust clean hypergraph for incomplete multi-view clustering
Chen, Y., Chen, H.-M., Xiang, B., Yuan, Z., Luo, C., Horng, S., & Li, T.-R. (2025) · Expert systems with applications · doi:10.1016/j.eswa.2025.130203 · Cited by 4
Chen, H., Gómez, C., Corrêa, Z. M., Liu, A., Milman, T., Eiger‐Moscovich, M., Chévez‐Barrios, P., Salomão, D. R., & Unberath, M. (2025) · Medical Image Analysis · doi:10.1016/j.media.2025.103856 · Cited by 3 · see citing works
Shao, C.-F., Zhao, Y.-S., Chen, Y.-S., Duan, J.-X., Zhang, Z.-Y., Wang, R., Liang, D., Chen, D.-Y., & Li, Z. (2026) · Engineering applications of artificial intelligence · doi:10.1016/j.engappai.2026.114461 · Cited by 3 · see citing works
Wang, H., Li, H., Ma, K.-R., Mo, G.-S., Yan, M.-H., Zhang, X.-Y., Xie, H.-N., Huang, Y.-Z., Li, H.-Y., Xue, Y.-W., et al. (2026) · npj Precision Oncology · doi:10.1038/s41698-026-01381-6 · Cited by 3 · see citing works
Non-small cell lung cancer subtype classification based on cross-scale multi-instance learning
Jiang, P., Chen, W., Zheng, G., Li, X., & Song, X. (2025) · Scientific Reports · doi:10.1038/s41598-025-27337-7 · Cited by 3 · see citing works
Meng, R., Zhu, W., Cameron, C. J. F., Ni, P.-Y., Zhou, X., Ulammandakh, T., & Gerstein, M. (2025) · Proceedings of the National Academy of Sciences of the United States of America · doi:10.1073/pnas.2423469122 · Cited by 3 · see citing works
Quantitative benchmarking of anomaly detection methods in digital pathology images
Cui, C., Zheng, X.-D., Deng, R.-N., Liu, Q., Yao, T.-Y., Wilson, K. T., Coburn, L. A., Landman, B. A., Yang, H.-C., Wang, Y.-H., et al. (2025) · Machine Learning. Health · doi:10.1088/3049-477x/ae0f9f · Cited by 3 · see citing works
Self-citationLin, Y., Li, Y., Long, X., Ye, Y., Guo, J., & Li, D. (2025) · Arabian Journal for Science and Engineering · doi:10.1007/s13369-025-10320-9 · Cited by 3 · see citing works
“In recent years, there has been a surge in the adoption of MIL [10–13] for the classification of WSI in tumor clas-sification.”
Retrosynthetic analysis via deep learning to improve pilomatricoma diagnoses
Wang, Z., Tan, X., Xue, Y., Hu, H., Lin, K., Wang, C., Fu, H., & Zhang, J. (2024) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2024.109152 · Cited by 3 · see citing works
Li, Y., Lin, Z.-H., Zhang, Q.-X., Ding, X., Yang, H., Pi, L.-J., Yuan, W., Wen, Y.-Q., Guo, L., Zhang, Q.-L., et al. (2026) · Medical Image Anal. · doi:10.1016/j.media.2026.104060 · Cited by 2 · see citing works
Liu, Z.-Q., Zhu, L., Chen, J., Lin, X., Xie, Y.-N., Zhao, J., & Ma, J.-Y. (2025) · Expert systems with applications · doi:10.1016/j.eswa.2025.130525 · Cited by 2
Poly Instance Recurrent Neural Network for Real-time Lifelong Learning at the Low-power Edge
Al-Ameen, S., Sudharsan, B., Vijayakumar, T., Szydło, T., Shah, T., & Ranjan, R. (2024) · BigData Congress [Services Society] · doi:10.1109/bigdata62323.2024.10825026 · Cited by 2
Multi-Scale Hierarchical Context-Aware Survival Prediction Network based on whole slide images
Song, J., Hao, Y., Zhai, H., Zhao, S., Tao, N. P., Huang, S., & Duan, X. (2025) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2025.108159 · Cited by 1 · see citing works
Retinal vessel junction detection using vessel guidance and heatmap regression
Long, J., Xie, Y., Que, N., & Yuan, P. (2025) · Measurement · doi:10.1016/j.measurement.2025.118966 · Cited by 1 · see citing works
Adaptive fusion with structural separation for incomplete multi-view clustering
Sun, L., Xiao, D., Long, X., Chen, D.-L., & Wang, Z. (2026) · Pattern Recognition · doi:10.1016/j.patcog.2026.114115 · Cited by 1
Interpretable multiple instance learning for hematologic diagnosis from peripheral blood smears
Singi, S., Sun, S.-H., Yin, Z.-H., Gupta, R., Webb, D. C., Bilal, K., Dilip, D., Wang, L.-L., Kumar, N., Nanda, S., et al. (2026) · Communications Medicine · doi:10.1038/s43856-026-01558-x · Cited by 1
“Multiple instance learning (MIL) offers a promising solution for weakly supervised learning in pathology 12,16 – 20 , treating individual cells as instances andthePBSasalabeledbag.”
2.5 CNN: Leveraging 2D CNNs to Pretrain 3D Models in Low-Data Regimes for COVID-19 Diagnosis
Garg, A., Garg, A., & Duncan, D. (2025) · Electronics · doi:10.3390/electronics14132571 · Cited by 1 · see citing works
Wu, X., Zhang, H., Mei, C., Wu, J., & Ai, H. (2025) · Symmetry · doi:10.3390/sym17040608
Yu, H., Wu, J., Xu, J., Wang, S., Wang, W., Chai, S., & Xin, J. (2024) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm62325.2024.10822836
“Our proposed method was evaluated against state-of-the-art WSI algorithms, including the attention-based AB-MIL [7], the CLAM [9], the DTFD-MIL [10], and the multi-scale CS-MIL [17].”
FORESEE: Multimodal and multiview representation learning for robust cancer survival prediction
Pan, L., Nie, C., Wang, X., & Peng, S. (2026) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2026.111196
Su, F. (2026) · Journal of Imaging Informatics in Medicine · doi:10.1007/s10278-026-02354-8
Gu, M.-K., Qi, Z., Chen, X.-Y., Zhang, D.-W., Sun, Y., Dang, C., & Qi, X. (2026) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2026.110936
Zhang, F.-Y., Hu, K.-M., Sun, C.-M., Liu, X.-F., Wei, L.-Y., & Su, R. (2026) · Medical Image Analysis · doi:10.1016/j.media.2026.104343
Mind the Gap: Continuous magnification sampling for pathology foundation models
Möllers, A., Hense, J., Schulz, F., Milbich, T., Alber, M., & Ruff, L. (2026) · Medical Image Analysis · doi:10.1016/j.media.2026.104327
Xu, N., Chen, A., Ruan, H., Yang, D.-L., Liao, R.-J., Qi, X., Fong, S., Cao, D.-D., Yu, L., Huang, Y.-H., et al. (2026) · PLOS Digital Health · doi:10.1371/journal.pdig.0001744
“Potentially, a multiscale design allowing more variable peri-LE context may also improve model robustness [57,66], as the physical boundary of the peri-LE stroma compartment has not yet been clearly defined by a fixed depth or anatomical cutoff [67,68].”
Huang, Y., Yuan, Y., Ye, W., & Ming, W. (2026) · Mathematics · doi:10.3390/math14173059
“Previous multi-scale pathology studies have shown the value of explicitly modelling inter-scale relationships across WSI magnifications [50].”
CSFMIL: Whole slide image classification with two-stage cross-scale fusion
Li, S., Luo, Y., Chen, Z.-N., Chen, F.-J., Hu, K., & Gao, X.-P. (2026) · Neurocomputing · doi:10.1016/j.neucom.2026.134807
Multimodal Artificial Intelligence in Tissue Diagnostics
Xia, R., Isett, B. R., Chen, J., Parwani, A. V., Bao, R., & Gu, Q. (2026) · American Journal Of Pathology · doi:10.1016/j.ajpath.2026.07.002 · Cited by 1 · see citing works
CLARISA: Connexin-43 Lateralization Automated ROI-Based Image Signal Analyzer
Gattari, D., Sancho-Zamora, J., Chan, D., Prado, N., Diez, E., Soria, M., & Rossi, M. (2026) · International Journal of Molecular Sciences · doi:10.3390/ijms27115033
“…detail to assess the local morphology of the CX43-positive region and its position within the myocardial fiber, whereas the larger crop provided contextual information on CX43 expression in neighboring fibers, following previous multi-scale designs proposed for medical image analysis [24,25].”
Ye, C., Li, G., Zeng, S., Hu, K., Lv, J., Zhou, Q., & Zhang, H. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11460571
Zhu, K.-R., & Samsudin, N.-H. (2026) · The Arabian journal for science and engineering · doi:10.1007/s13369-026-11293-z
“…morphologies that conventional CNNs fail to model [1]; (2) Low-contrast tissue interfaces obscure boundaries where pure attention mechanisms lose spatial fidelity [2]; and (3) Multi-scale pathology presentations demand adaptive feature integration across orders of magnitude in lesion size [3].”
Zhu, Z., Meijering, E. H. W., & Wang, L. S. (2026) · Medical Image Analysis · doi:10.1016/j.media.2026.104082
A Multi-Scale Linear-Time Encoder for Whole-Slide Image Analysis
Dwarampudi, J., Wong, J., Nguyen, H. V., & Banerjee, T. (2026) · IEEE International Symposium on Biomedical Imaging · doi:10.1109/isbi61048.2026.11516046
“Multi-scale MIL fused magnifications through hierarchical attention or semantic filtering (for example HAG-MIL and CS-MIL) [4, 5].”
Mahmood, K., Althobaiti, M. M., Hassan, M. U., Khan, S., Khan, M., & Alsoufi, M. A. (2026) · Scientific Reports · doi:10.1038/s41598-026-36622-y
Enhancing whole slide image classification via fine-grained adaptive scale fusion
Yu, J., Wang, Z., Sun, S., Wu, B., Zhou, S., Zhu, J., Hu, L., Wu, Y., & Han, S. (2026) · Signal Image and Video Processing · doi:10.1007/s11760-025-05065-9
“In studies using dynamic weights [33, 34], while this issue is avoided, the use of coarse-grained fusion and the lack of further refinement for better-performing scales still limit the development of scale fusion methods.”
Lightweight Breast Histopathological Classification Model Based on Dynamic Multi-Scale Selection
Hong, G., & 彭文柱 (2025) · International Symposium on AI-Driven Image Processing and Computer Vision Technology (AIPCVT) · doi:10.1109/aipcvt67868.2025.11405634
“Drawing on the work of Marini et al. [10] and Deng et al. [11], LightMIL also places particular emphasis on enhancing the model's ability to identify key lesion regions under weakly supervised conditions, thereby achieving more stable and reliable classification performance for breast…”
WaveConvX: Multi-Level Wavelet Enhancement for Histopathology Image Classification
Zhang, Y.-Z., Fang, Y., Liu, F.-R., Liu, Z.-X., & Jia, X.-B. (2025) · IEEE International Conference on Systems, Man and Cybernetics · doi:10.1109/smc58881.2025.11343022
“Nonetheless, pathological images contain subtle morphological cues spanning multiple scales [8], [20], and naive CNN-based pipelines may lose some high-frequency detail during aggressive downsampling.”
Yue, L., Li, H., & Huang, W. (2025) · 2025 8th International Conference on Computer Information Science and Application Technology (CISAT) · doi:10.1109/cisat66811.2025.11181949
“Approaches include: domain adversarial networks for cancer subtyping [5], attention-guided frameworks for survival analysis [6,7], and GCNs modeling inter-patch relationships in WSIs [8].”
Patch Target Guided Dual-Branch Deep Multiple Instance Learning for 3D MRI Analysis
Dai, M., Shi, X., Zhu, X., Pan, T., & Li, K. (2024) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm62325.2024.10821985 · Cited by 3 · see citing works
- Authors
- Deng, R., Cui, C., Remedios, L. W., Bao, S., Womick, R. M., Chiron, S., Li, J., Roland, J. T., Lau, K. S., Liu, Q., Wilson, K. T., Wang, Y., Coburn, L. A., Landman, B. A., & Huo, Y.
- Affiliations
- Vanderbilt University (US)
- University of North Carolina at Chapel Hill (US)
- Vanderbilt University Medical Center (US)
- VA Tennessee Valley Healthcare System (US)
- Corresponding author
- Huo, Yuankai
- ORCID
- Venue
- Medical Image Analysis
- Publisher
- Elsevier BV
- Published
- 2024-05
- Volume
- 94
- Pages
- 103124
- ISSN
- 1361-8415 (ISSN-L)
- PMC
- PMC11016375
- OpenAlex
- W4392190141
- Cited by
- 95
- References
- 41
- Citations by year
- 2026: 34 · 2025: 44 · 2024: 15 · 2023: 2
- FWCI
- 13.72
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- Locations
- Medical Image Analysis (publishedVersion, cc-by)
- PubMed (publishedVersion)
- PubMed Central (acceptedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- NIH (R01DK135597)
- U.S. Department of Veterans Affairs (I01CX002171, I01BX004366)
- National Science Foundation (1452485)
- Leona M and Harry B Helmsley Charitable Trust (G-1903-03793)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › AI in cancer detection
- Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › Radiomics and Machine Learning in Medical Imaging
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Image Retrieval and Classification Techniques
- Keywords
- whole slide images, digital pathology, multiscale image analysis, cross-scale attention
- MeSH terms
- Algorithms, Diagnostic Imaging, Humans
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
