Computer Methods and Programs in Biomedicine, 2024

[Journal Link] | [arXiv] | [Related Resources] | [Citation]

Abstract: (1) Methods: This scheme first designs an attention-based method to generate a bag prototype for each slide. On this basis, it further groups WSI patch instances into a series of instance clusters according to the feature similarities between the prototype and patches. Finally, pseudo-bags are obtained by randomly combining the non-overlapping patch instances of different instance clusters. Moreover, the design scheme of our ProDiv considers practicality, and it could be smoothly assembled with almost all the MIL-based WSI classification methods in recent years. (2) Conclusions: ProDiv could almost always bring obvious performance improvements to compared MIL models on typical metrics, which suggests the effectiveness of our scheme. Experimental visualization also visually interprets the correctness of the proposed ProDiv.

We use the pre-trained ResNet50 for extracting patch features. You can download the pre-trained ResNet50 at here.

An improved version (ProDiv+) can be found at PseMix. Moreover, ProDiv is also contained. For more details, you can move to the PseMix repo.

Here we list the related works involving pseudo-bags or using pseudo-bags for training deep MIL networks.

If you find this work helps your research, please consider citing our paper via

@article{YANG2024108161,
    author = {Yang, Rui and Liu, Pei and Ji, Luping},
    doi = {https://doi.org/10.1016/j.cmpb.2024.108161},
    issn = {0169-2607},
    journal = {Computer Methods and Programs in Biomedicine},
    pages = {108161},
    title = {{ProDiv: Prototype-driven consistent pseudo-bag division for whole-slide image classification}},
    url = {https://www.sciencedirect.com/science/article/pii/S0169260724001573},
    volume = {249},
    year = {2024}
}