Deep Adversarial Training for Multi-Organ Nuclei Segmentation in Histopathology Images

By Faisal Mahmood, Daniel Borders, Richard J. Chen, Gregory N. McKay, Kevan J. Salimian, A. Baras, Nicholas J. Durr. IEEE Transactions on Medical Imaging, 2018

331 citations recorded by the source corpus.

Nuclei mymargin segmentation is a fundamental task for various computational pathology applications including nuclei morphology analysis, cell type classification, and cancer grading. Deep learning has emerged as a powerful approach to segmenting nuclei but the accuracy of convolutional neural networks (CNNs) depends on the volume and the quality of labeled histopathology data for training. In particular, conventional CNN-based approaches lack structured prediction capabilities, which are required to distinguish overlapping and clumped nuclei. Here, we present an approach to nuclei segmentation that overcomes these challenges by utilizing a conditional generative adversarial network (cGAN) trained with synthetic and real data. We generate a large dataset of H&E training images with perfect nuclei segmentation labels using an unpaired GAN framework. This synthetic data along with real histopathology data from six different organs are used to train a conditional GAN with spectral normalization and gradient penalty for nuclei segmentation. This adversarial regression framework enforces higher-order spacial-consistency when compared to conventional CNN models. We demonstrate that this nuclei segmentation approach generalizes across different organs, sites, patients and disease states, and outperforms conventional approaches, especially in isolating individual and overlapping nuclei.

Subfields: Artificial Intelligence, Biophysics, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/7ae5b1b4b488473314d40711e4d2b0e8d7e210ed

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Biophysics

In Computer Vision and Pattern Recognition

Citations in the corpus

Scientific discovery in the age of artificial intelligence (2023, background): “…for scientific images because they synthesize realistic images in many domains ranging from particle collision events 42 , pathology slides 43 , chest X-rays 44 , magnetic resonance contrasts 45 , three-dimensional (3D) material microstructure 46 , protein functions 47,48 to genetic…”
Artificial intelligence and illusions of understanding in scientific research (2024, unknown): “In the physical and biological sciences, proposed applications for generative AI include augmenting small datasets or simulating new data to study phenomena such as cosmological structures 42 , medical images 43,44 and nucleic-acid sequences 45 .”
Multimodal Co-Attention Transformer for Survival Prediction in Gigapixel Whole Slide Images (2021, background): “…the entire composition of a WSI and dependent on the interactions of its components ( i.e. - a ﬁne-grained visual recognition problem such as one that involves understanding the complex milieu of stroma, tumor aggregates, immune cells and other visual concepts) [68, 7, 47, 22, 40, 21, 6, 39, 15].”
Methods for Segmentation and Classification of Digital Microscopy Tissue Images (2018, background): “While there are some recent approaches for generation of synthetic datasets (e.g., Hou et al., 2017; Mahmood et al., 2018), manually annotated datasets continue to represent gold standard data for training and testing, and the paucity of large, manually annotated datasets remains to be a reality.”
A Multi-Organ Nucleus Segmentation Challenge (2020, methodology): “Deep generative models have also been used for accurate nuclei segmentation [25].”
SynthSeg: Segmentation of brain MRI scans of any contrast and resolution without retraining (2021, unknown): “Synthetic training data can be used to increase robustness by introducing surrogate domain variations, either generated with physics-based models [42], or adversarial generative networks [26, 10], possibly conditioned on label maps for improved semantic content [48, 40].”
Cancer Diagnosis Using Deep Learning: A Bibliographic Review (2019, methodology): “Goodfellow et al. introduced a method named as a generative adversarial network GAN [110], which is basically a two-player min-max game having a generator as the ﬁrst player and discriminatory as the second player.”