# Semi-Supervised Histology Classification using Deep Multiple Instance Learning and Contrastive Predictive Coding

Source: [Semi-Supervised Histology Classification using Deep Multiple Instance Learning and Contrastive Predictive Coding](<https://www.sota2.com/research/paper/68fa5ef41e209131ce9e68a9>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-25T19:04:52.118Z

Authors: Ming Y. Lu, Richard J. Chen, Jingwen Wang, Debora Dillon, Faisal Mahmood

Published: 2019-10-23T22:12:57.000Z

## Abstract

Convolutional neural networks can be trained to perform histology slide classification using weak annotations with multiple instance learning (MIL). However, given the paucity of labeled histology data, direct application of MIL can easily suffer from overfitting and the network is unable to learn rich feature representations due to the weak supervisory signal. We propose to overcome such limitations with a two-stage semi-supervised approach that combines the power of data-efficient self-supervised feature learning via contrastive predictive coding (CPC) and the interpretability and flexibility of regularized attention-based MIL. We apply our two-stage CPC + MIL semi-supervised pipeline to the binary classification of breast cancer histology images. Across five random splits, we report state-of-the-art performance with a mean validation accuracy of 95% and an area under the ROC curve of 0.968. We further evaluate the quality of features learned via CPC relative to simple transfer learning and show that strong classification performance using CPC features can be efficiently leveraged under the MIL framework even with the feature encoder frozen.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Lymphoma Tissue Classification on Janowczyk (test)](<https://www.sota2.com/research/sota/lymphoma-tissue-classification-on-janowczyk-test.md>): Lymphoma Tissue Classification; Janowczyk (test); Accuracy: 45.2.

- [Gleason Scoring Classification on Arvaniti (test)](<https://www.sota2.com/research/sota/gleason-scoring-classification-on-arvaniti-test.md>): Gleason Scoring Classification; Arvaniti (test); Accuracy: 68.1.

- [Breast Tissue Classification on Litjens (test)](<https://www.sota2.com/research/sota/breast-tissue-classification-on-litjens-test.md>): Breast Tissue Classification; Litjens (test); Accuracy: 0.732.

- [Head & Neck Tissue Classification on Shaban (test)](<https://www.sota2.com/research/sota/head-and-neck-tissue-classification-on-shaban-test.md>): Head & Neck Tissue Classification; Shaban (test); Accuracy: 68.3.

- [Lung Tissue Classification on Alsubaie (test)](<https://www.sota2.com/research/sota/lung-tissue-classification-on-alsubaie-test.md>): Lung Tissue Classification; Alsubaie (test); Accuracy: 73.4.

- [Meningioma Tissue Classification on Qureshi (test)](<https://www.sota2.com/research/sota/meningioma-tissue-classification-on-qureshi-test.md>): Meningioma Tissue Classification; Qureshi (test); Accuracy: 56.6.

- [Multi-label Tissue Classification on Hosseini et al. (test)](<https://www.sota2.com/research/sota/multi-label-tissue-classification-on-hosseini-et-al-test.md>): Multi-label Tissue Classification; Hosseini et al. (test); Avg Binary Accuracy: 0.775.

- [Colon Tissue Classification on Kather (test)](<https://www.sota2.com/research/sota/colon-tissue-classification-on-kather-test.md>): Colon Tissue Classification; Kather (test); Accuracy: 77.4.

- [Nuclear Atypia Classification on Roux (test)](<https://www.sota2.com/research/sota/nuclear-atypia-classification-on-roux-test.md>): Nuclear Atypia Classification; Roux (test); Accuracy: 66.1.

- [Ovary Tissue Classification on Kobel (test)](<https://www.sota2.com/research/sota/ovary-tissue-classification-on-kobel-test.md>): Ovary Tissue Classification; Kobel (test); Accuracy: 65.1.

- [Nuclear Mitosis Classification on Veta (test)](<https://www.sota2.com/research/sota/nuclear-mitosis-classification-on-veta-test.md>): Nuclear Mitosis Classification; Veta (test); Accuracy: 57.5.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/1910.10825>)

[PDF](<https://arxiv.org/pdf/1910.10825.pdf>)
