Abstract
The current study of cell architecture of inflammation in histopathology images commonly performed for diagnosis and research purposes excludes a lot of information available on the biopsy slide. In autoimmune diseases, major outstanding research questions remain regarding which cell types participate in inflammation at the tissue level, and how they interact with each other. While these questions can be partially answered using traditional methods, artificial intelligence approaches for segmentation and classification provide a much more efficient method to understand the architecture of inflammation in autoimmune disease, holding a great promise for novel insights. In this paper, we empirically develop deep learning approaches that uses dermatomyositis biopsies of human tissue to detect and identify inflammatory cells. Our approach improves classification performance by 26% and segmentation performance by 5%. We also propose a novel post-processing autoencoder architecture that improves segmentation performance by an additional 3%.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
References
Agarwal, V., Jhalani, H., Singh, P., Dixit, R.: Classification of melanoma using efficient nets with multiple ensembles and metadata. In: Tiwari, R., Mishra, A., Yadav, N., Pavone, M. (eds.) Proceedings of International Conference on Computational Intelligence. AIS, pp. 101–111. Springer, Singapore (2022). https://doi.org/10.1007/978-981-16-3802-2_8
Brock, A., De, S., Smith, S.L., Simonyan, K.: High-performance large-scale image recognition without normalization. In: International Conference on Machine Learning, pp. 1059–1071. PMLR (2021)
Dash, M., Londhe, N.D., Ghosh, S., Semwal, A., Sonawane, R.S.: Pslsnet: Automated psoriasis skin lesion segmentation using modified u-net-based fully convolutional network. Biomed. Signal Process. Contr. 52 226–237 (2019). https://doi.org/10.1016/j.bspc.2019.04.002, https://www.sciencedirect.com/science/article/pii/S1746809419300990
Dinse, G.E., et al.: Increasing prevalence of antinuclear antibodies in the united states. Arthritis Rheumatol. 72(6), 1026–1035 (2020)
Dosovitskiy, A., et al.: An image is worth 16x16 words: Transformers for image recognition at scale. ar**v preprint ar**v:2010.11929 (2020)
Ehrenfeld, M., et al.: Covid-19 and autoimmunity. Autoimmun. Rev. 19(8), 102597 (2020)
Falcon, W., et al.: Pytorch lightning. GitHub. Note: https://github.com/PyTorchLightning/pytorch-lightning vol. 3(6) (2019)
Galeotti, C., Bayry, J.: Autoimmune and inflammatory diseases following covid-19. Nat. Rev. Rheumatol. 16(8), 413–414 (2020)
He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learningfor image recognition. In: ComputerScience (2015)
Hu, J., Shen, L., Sun, G.: Squeeze-and-excitation networks. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 7132–7141 (2018)
Izmailov, P., Podoprikhin, D., Garipov, T., Vetrov, D., Wilson, A.G.: Averaging weights leads to wider optima and better generalization. ar**v preprint ar**v:1803.05407 (2018)
Jacobson, D.L., Gange, S.J., Rose, N.R., Graham, N.M.: Epidemiology and estimated population burden of selected autoimmune diseases in the united states. Clin. Immunol. Immunopathol. 84(3), 223–243 (1997)
Lerner, A., Jeremias, P., Matthias, T.: The world incidence and prevalence of autoimmune diseases is increasing. Int. J. Celiac Disease 3(4), 151–155 (2015). 10.12691/ijcd-3-4-8, http://pubs.sciepub.com/ijcd/3/4/8
Lin, T.Y., Goyal, P., Girshick, R., He, K., Dollár, P.: Focal loss for dense object detection. In: Proceedings of the IEEE International Conference on Computer Vision, pp. 2980–2988 (2017)
Liu, Z., et al.: Swin transformer: Hierarchical vision transformer using shifted windows. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 10012–10022 (2021)
Liu, Z., Mao, H., Wu, C.Y., Feichtenhofer, C., Darrell, T., **e, S.: A convnet for the 2020s. ar**v preprint ar**v:2201.03545 (2022)
Picard, D.: Torch.manual_seed(3407) is all you need: On the influence of random seeds in deep learning architectures for computer vision. CoRR abs/2109.08203 (2021). https://arxiv.org/abs/2109.08203
Raghu, M., Zhang, C., Kleinberg, J., Bengio, S.: Transfusion: Understanding transfer learning for medical imaging. In: Advances in Neural Information Processing Systems, vol. 32 (2019)
Ronneberger, O., Fischer, P., Brox, T.: U-Net: convolutional networks for biomedical image segmentation. In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F. (eds.) MICCAI 2015. LNCS, vol. 9351, pp. 234–241. Springer, Cham (2015). https://doi.org/10.1007/978-3-319-24574-4_28
Stafford, I., Kellermann, M., Mossotto, E., Beattie, R., MacArthur, B., Ennis, S.: A systematic review of the applications of artificial intelligence and machine learning in autoimmune diseases. NPJ Digital Med. 3(1), 1–11 (2020)
Tan, M., Le, Q.: Efficientnet: Rethinking model scaling for convolutional neural networks. In: International Conference on Machine Learning, pp. 6105–6114. PMLR (2019)
Tsakalidou, V.N., Mitsou, P., Papakostas, G.A.: Computer vision in autoimmune diseases diagnosis—current status and perspectives. In: Smys, S., Tavares, J.M.R.S., Balas, V.E. (eds.) Computational Vision and Bio-Inspired Computing. AISC, vol. 1420, pp. 571–586. Springer, Singapore (2022). https://doi.org/10.1007/978-981-16-9573-5_41
Buren, V., et al.: Artificial intelligence and deep learning to map immune cell types in inflamed human tissue. Journal of Immunological Methods 505, 113233 (2022). https://doi.org/10.1016/j.jim.2022.113233, https://www.sciencedirect.com/science/article/pii/S0022175922000205
Wightman, R.: Pytorch image models. https://github.com/rwightman/pytorch-image-models (2019). https://doi.org/10.5281/zenodo.4414861
**e, C., Tan, M., Gong, B., Wang, J., Yuille, A.L., Le, Q.V.: Adversarial examples improve image recognition. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 819–828 (2020)
**e, Q., Luong, M.T., Hovy, E., Le, Q.V.: Self-training with noisy student improves imagenet classification. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 10687–10698 (2020)
Zhou, Z., Rahman Siddiquee, M.M., Tajbakhsh, N., Liang, J.: UNet++: a nested u-net architecture for medical image segmentation. In: Stoyanov, D., et al. (eds.) DLMIA/ML-CDS -2018. LNCS, vol. 11045, pp. 3–11. Springer, Cham (2018). https://doi.org/10.1007/978-3-030-00889-5_1
Acknowledgment
We would like to thank NYU HPC team for assisting us with our computational needs. We would also like to thank Prof. Elena Sizikova (Moore Sloan Faculty Fellow, Center for Data Science (CDS), New York University (NYU)) for her valuable feedback.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
A Appendix
A Appendix
1.1 A.1 Expansion of Results
In Tables 6 and 7 we show complete results with mean and standard deviation. These are an expansion of Table 2 in Sect. 5.1 of the main paper. Tables were compressed to save space and only focus on the main results. To provide a complete picture, we added extended results in this section.
1.2 A.2 Autoencoder with Efficientnet Encoder for Segmentation
In Tables 8 and 9 we compared the time taken to train and the performance of the respective trained architecture for segmentation using EfficientNet encoders. We observed that with the addition of autoencoder post-processing, training time increased by an average of 3 m 7.3 s over 50 epochs (averaged over the entire efficientnet family). This is an increase of 2.93% in training time over the eight encoders. In other words, an average increase of 0.36% increase in time per encoder over 50 epochs.
Performance wise architecture with autoencoder post-processing consistently outperformed segmentation architectures without them by 2.75%.
Similarly, we compared computational and performance for UNet++ with and without the autoencoder post-processing in Tables 10 and 11 respectively. In this case, we observed that the gain in performance with autoencoder post-processing is 5% averaged over the efficientnet family of encoders. This also corresponds to a 3 m 7 s increase in training time which is an increase of 2.6%.
1.3 A.3 Metrics Description
For measuring segmentation performance, we use IoU or intersection over union metric. It helps us understand how similar sample sets are.
![figure a](http://media.springernature.com/lw685/springer-static/image/chp%3A10.1007%2F978-3-031-25066-8_21/MediaObjects/544157_1_En_21_Figa_HTML.png)
Here the comparison is made between the output mask by segmentation pipeline against the ground truth mask.
For measuring classification performance, we use the F1 score.
Computed as F1 = \(\frac{\text {2*Precision*Recall}}{\text {Precision+Recall}} = \frac{\text {2*TP}}{\text {2*TP+FP+FN}}\)
1.4 A.4 Effect of Different Weights
ImageNet initialization has been the defacto norm for most transfer learning tasks. Although in some cases, as in [1] it was observed that noisy student weights performed better than ImageNet initialization. To study the effect in our case, we used advprop and noisy student initialization. ImageNet weights for initialization work for medical data not because of feature reuse but because of better weight scaling and faster convergence [18]. Noisy student training [26] extends the idea of self-training and distillation with the use of equal-or-larger student models, and noise such as dropout, stochastic depth, and data augmentation via RandAugment is added to the student during learning so that the student generalizes better than the teacher. First, an EfficientNet model is trained on labelled images and is used as a teacher to generate pseudo labels for 300M unlabeled images. We then train a larger EfficientNet as a student model on the combination of labelled and pseudo-labelled images. This helps reduce the error rate, increases robustness and improves performance over the existing state-of-the-art on ImageNet.
(ii)AdvProp training, which banks on Adversarial examples, which are commonly viewed as a threat to ConvNets. In [25] they present an opposite perspective: adversarial examples can be used to improve image recognition models. They treat adversarial examples as additional examples to prevent overfitting. It performs better when the models are bigger. This improves upon performance for various ImageNet and its’ subset benchmarks.
Since initially all these were developed for the EfficientNet family of the encoders, we used them for benchmarking. We present their results in Table 12.
Similarly, we conduct similar experiments for classification with different initialization. We reported these results in Table 13.
As we can see for segmentation, ImageNet initialization performed better in most cases. Similarly, in classification, it not only performed better in most cases but also provided the best overall result—these inferences, combined with the fact that advprop and noisy student requires additional computational resources. Hence we decide to stick with ImageNet initialization.
1.5 A.5 Expansion on Experimental Details
Segmentation. We used PyTorch lightning’s [7] seed everything functionality to seed all the generator values uniformly. For setting the seed value, we randomly generated a set of numbers in the range of 1 and 1000. We did not perform an extensive search of space to optimise performance with seed value as suggested in [17]. We used seed values 26, 77, 334, 517 and 994. For augmentation, we used conversion to PIL Image to apply random rotation (degrees=3), random vertical and horizontal flip, then conversion to tensor and finally channel normalisation. We could have used a resize function to reshape the 1408 by 1876 Whole Slide Images (WSI), but we instead tilled them in 480 square tile images. We then split them into a batch size of 16 before finally passing through the segmentation architecture (UNet/UNet++). We used channel attention only decoder, with ImageNet initialisation and a decoder depth of 3 (256, 128, 64).
We used cross-entropy loss with dark/light pixel normalization, Adam optimizer with LR set to 3.6e-04 and weight decay of 1e-05. We used a cosine scheduling rate with a minimum value set to 3.4e-04.
APP Segmentation. When using APP we used GELU activation by default with adam optimizer and lr set to 1e-3.
Classification. For Classification, we used the same seed values with PyTorch lightning’s [7] seed everything functionality, as described for segmentation above. For augmentation, we resized the images to 384 square images, followed by randomly applying colour jitter (0.2, 0.2, 0.2) or random perspective (distortion scale=0.2) with probability 0.3, colour jittering (0.2, 0.2, 0.2) or random affine (degrees=10) with probability 0.3, random vertical flip and random horizontal flip with probability 0.3 and finally channel normalization.
We used Stochastic weigh averaging with adam optimizer. We used a cosine learning rate starting at 1e-3 and a minimum set to 1e-6. We used focal loss with normalized class weight as our loss function. We used 6-fold validation with each fold of 20 epochs and batch size of 16. We used same parameters for both CNN and Transformers.
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Singh, P., Cirrone, J. (2023). A Data-Efficient Deep Learning Framework for Segmentation and Classification of Histopathology Images. In: Karlinsky, L., Michaeli, T., Nishino, K. (eds) Computer Vision – ECCV 2022 Workshops. ECCV 2022. Lecture Notes in Computer Science, vol 13803. Springer, Cham. https://doi.org/10.1007/978-3-031-25066-8_21
Download citation
DOI: https://doi.org/10.1007/978-3-031-25066-8_21
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-25065-1
Online ISBN: 978-3-031-25066-8
eBook Packages: Computer ScienceComputer Science (R0)