Michael Yeung
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Michael Yeung has a total of 46 co-authors. Here are the top co-authors by number of shared publications:
- Evis Sala
Dipartimento di Scienze Radiologiche ed Ematologiche, Università Cattolica del Sacro Cuore, Rome, Italy
6 shared publications - Carola-Bibiane Schönlieb
University of Cambridge, UK
5 shared publications - Leonardo Rundo
Department of Information Engineering, Electrical Engineering and Applied Mathematics (DIEM), University of Salerno, 84084 Fisciano, Italy
5 shared publications - Guang Yang
Bioengineering Department and Imperial-X, Imperial College London, London, UK
4 shared publications - Michael J. Hendrickson
University of North Carolina Chapel Hill, Chapel Hill, NC
3 shared publications - Sameer Arora
The University of North Carolina at Chapel Hill, United States
3 shared publications
Latest Publications
- Open Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of Engineering in Medicine and Biology
- Vol. 7, 187-195
Stain Consistency Learning: Handling Stain Variation for Automatic Digital Pathology Segmentation
Stain variation poses a major challenge for automated digital pathology. Numerous techniques address this issue, yet show limited success, especially outside H&E stains and classification tasks. We propose Stain Consistency Learning (SCL), combining stain-specific augmentation and a novel consistency loss to learn stain-invariant features. We conduct the first large-scale evaluation of ten methods on Massons trichrome and H&E datasets for segmentation. Our results demonstrate that traditional stain normalization offers little benefit, while stain augmentation and adversarial learning significantly improve performance. SCL consistently outperforms all other methods.
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of Engineering in Medicine and Biology
- Vol. 7, 187-195
Open AccessStain Consistency Learning: Handling Stain Variation for Automatic Digital Pathology Segmentation
Stain variation poses a major challenge for automated digital pathology. Numerous techniques address this issue, yet show limited success, especially outside H&E stains and classification tasks. We propose Stain Consistency Learning (SCL), combining stain-specific augmentation and a novel consistency loss to learn stain-invariant features. We conduct the first large-scale evaluation of ten methods on Massons trichrome and H&E datasets for segmentation. Our results demonstrate that traditional stain normalization offers little benefit, while stain augmentation and adversarial learning significantly improve performance. SCL consistently outperforms all other methods.
- Open Access
Stain Consistency Learning: Handling Stain Variation for Automatic Digital Pathology Segmentation
Stain variation is a unique challenge associated with automated analysis of digital pathology. Numerous methods have been developed to improve the robustness of machine learning methods to stain variation, but comparative studies have demonstrated limited benefits to performance. Moreover, methods to handle stain variation were largely developed for H&E stained data, with evaluation generally limited to classification tasks. Here we propose Stain Consistency Learning, a novel framework combining stain-specific augmentation with a stain consistency loss function to learn stain colour invariant features. We perform the first, extensive comparison of methods to handle stain variation for segmentation tasks, comparing ten methods on Masson's trichrome and H&E stained cell and nuclei datasets, respectively. We observed that stain normalisation methods resulted in equivalent or worse performance, while stain augmentation or stain adversarial methods demonstrated improved performance, with the best performance consistently achieved by our proposed approach. The code is available at: https://github.com/mlyg/stain_consistency_learning
Open AccessStain Consistency Learning: Handling Stain Variation for Automatic Digital Pathology Segmentation
Stain variation is a unique challenge associated with automated analysis of digital pathology. Numerous methods have been developed to improve the robustness of machine learning methods to stain variation, but comparative studies have demonstrated limited benefits to performance. Moreover, methods to handle stain variation were largely developed for H&E stained data, with evaluation generally limited to classification tasks. Here we propose Stain Consistency Learning, a novel framework combining stain-specific augmentation with a stain consistency loss function to learn stain colour invariant features. We perform the first, extensive comparison of methods to handle stain variation for segmentation tasks, comparing ten methods on Masson's trichrome and H&E stained cell and nuclei datasets, respectively. We observed that stain normalisation methods resulted in equivalent or worse performance, while stain augmentation or stain adversarial methods demonstrated improved performance, with the best performance consistently achieved by our proposed approach. The code is available at: https://github.com/mlyg/stain_consistency_learning
- Open Access
- 6 December 2022
- Published by Springer Nature in Journal of Digital Imaging
- Vol. 36 (2) , 739-752
Calibrating the Dice Loss to Handle Neural Network Overconfidence for Biomedical Image Segmentation
The Dice similarity coefficient (DSC) is both a widely used metric and loss function for biomedical image segmentation due to its robustness to class imbalance. However, it is well known that the DSC loss is poorly calibrated, resulting in overconfident predictions that cannot be usefully interpreted in biomedical and clinical practice. Performance is often the only metric used to evaluate segmentations produced by deep neural networks, and calibration is often neglected. However, calibration is important for translation into biomedical and clinical practice, providing crucial contextual information to model predictions for interpretation by scientists and clinicians. In this study, we provide a simple yet effective extension of the DSC loss, named the DSC++ loss, that selectively modulates the penalty associated with overconfident, incorrect predictions. As a standalone loss function, the DSC++ loss achieves significantly improved calibration over the conventional DSC loss across six well-validated open-source biomedical imaging datasets, including both 2D binary and 3D multi-class segmentation tasks. Similarly, we observe significantly improved calibration when integrating the DSC++ loss into four DSC-based loss functions. Finally, we use softmax thresholding to illustrate that well calibrated outputs enable tailoring of recall-precision bias, which is an important post-processing technique to adapt the model predictions to suit the biomedical or clinical task. The DSC++ loss overcomes the major limitation of the DSC loss, providing a suitable loss function for training deep learning segmentation models for use in biomedical and clinical practice. Source code is available at https://github.com/mlyg/DicePlusPlus.
- 6 December 2022
- Published by Springer Nature in Journal of Digital Imaging
- Vol. 36 (2) , 739-752
Open AccessCalibrating the Dice Loss to Handle Neural Network Overconfidence for Biomedical Image Segmentation
The Dice similarity coefficient (DSC) is both a widely used metric and loss function for biomedical image segmentation due to its robustness to class imbalance. However, it is well known that the DSC loss is poorly calibrated, resulting in overconfident predictions that cannot be usefully interpreted in biomedical and clinical practice. Performance is often the only metric used to evaluate segmentations produced by deep neural networks, and calibration is often neglected. However, calibration is important for translation into biomedical and clinical practice, providing crucial contextual information to model predictions for interpretation by scientists and clinicians. In this study, we provide a simple yet effective extension of the DSC loss, named the DSC++ loss, that selectively modulates the penalty associated with overconfident, incorrect predictions. As a standalone loss function, the DSC++ loss achieves significantly improved calibration over the conventional DSC loss across six well-validated open-source biomedical imaging datasets, including both 2D binary and 3D multi-class segmentation tasks. Similarly, we observe significantly improved calibration when integrating the DSC++ loss into four DSC-based loss functions. Finally, we use softmax thresholding to illustrate that well calibrated outputs enable tailoring of recall-precision bias, which is an important post-processing technique to adapt the model predictions to suit the biomedical or clinical task. The DSC++ loss overcomes the major limitation of the DSC loss, providing a suitable loss function for training deep learning segmentation models for use in biomedical and clinical practice. Source code is available at https://github.com/mlyg/DicePlusPlus.
- Open Access
Calibrating the Dice loss to handle neural network overconfidence for biomedical image segmentation
The Dice similarity coefficient (DSC) is both a widely used metric and loss function for biomedical image segmentation due to its robustness to class imbalance. However, it is well known that the DSC loss is poorly calibrated, resulting in overconfident predictions that cannot be usefully interpreted in biomedical and clinical practice. Performance is often the only metric used to evaluate segmentations produced by deep neural networks, and calibration is often neglected. However, calibration is important for translation into biomedical and clinical practice, providing crucial contextual information to model predictions for interpretation by scientists and clinicians. In this study, we provide a simple yet effective extension of the DSC loss, named the DSC++ loss, that selectively modulates the penalty associated with overconfident, incorrect predictions. As a standalone loss function, the DSC++ loss achieves significantly improved calibration over the conventional DSC loss across six well-validated open-source biomedical imaging datasets, including both 2D binary and 3D multi-class segmentation tasks. Similarly, we observe significantly improved calibration when integrating the DSC++ loss into four DSC-based loss functions. Finally, we use softmax thresholding to illustrate that well calibrated outputs enable tailoring of recall-precision bias, which is an important post-processing technique to adapt the model predictions to suit the biomedical or clinical task. The DSC++ loss overcomes the major limitation of the DSC loss, providing a suitable loss function for training deep learning segmentation models for use in biomedical and clinical practice. Source code is available at: https://github.com/mlyg/DicePlusPlus.
Open AccessCalibrating the Dice loss to handle neural network overconfidence for biomedical image segmentation
The Dice similarity coefficient (DSC) is both a widely used metric and loss function for biomedical image segmentation due to its robustness to class imbalance. However, it is well known that the DSC loss is poorly calibrated, resulting in overconfident predictions that cannot be usefully interpreted in biomedical and clinical practice. Performance is often the only metric used to evaluate segmentations produced by deep neural networks, and calibration is often neglected. However, calibration is important for translation into biomedical and clinical practice, providing crucial contextual information to model predictions for interpretation by scientists and clinicians. In this study, we provide a simple yet effective extension of the DSC loss, named the DSC++ loss, that selectively modulates the penalty associated with overconfident, incorrect predictions. As a standalone loss function, the DSC++ loss achieves significantly improved calibration over the conventional DSC loss across six well-validated open-source biomedical imaging datasets, including both 2D binary and 3D multi-class segmentation tasks. Similarly, we observe significantly improved calibration when integrating the DSC++ loss into four DSC-based loss functions. Finally, we use softmax thresholding to illustrate that well calibrated outputs enable tailoring of recall-precision bias, which is an important post-processing technique to adapt the model predictions to suit the biomedical or clinical task. The DSC++ loss overcomes the major limitation of the DSC loss, providing a suitable loss function for training deep learning segmentation models for use in biomedical and clinical practice. Source code is available at: https://github.com/mlyg/DicePlusPlus.
- 25 July 2022
- Published by Springer Nature
- p. 547-561
From Astronomy to Histology: Adapting the FellWalker Algorithm to Deep Nuclear Instance Segmentation
- 25 July 2022
- Published by Springer Nature
- p. 547-561
From Astronomy to Histology: Adapting the FellWalker Algorithm to Deep Nuclear Instance Segmentation
- Top Cited PapersOpen Access
- 1 January 2022
- Published by Elsevier in Computerized Medical Imaging and Graphics
- Vol. 95, 102026
Unified Focal loss: Generalising Dice and cross entropy-based losses to handle class imbalanced medical image segmentation
- 1 January 2022
- Published by Elsevier in Computerized Medical Imaging and Graphics
- Vol. 95, 102026
Top Cited PapersOpen AccessUnified Focal loss: Generalising Dice and cross entropy-based losses to handle class imbalanced medical image segmentation
- Open Access
Unified Focal loss: Generalising Dice and cross entropy-based losses to handle class imbalanced medical image segmentation
Automatic segmentation methods are an important advancement in medical image analysis. Machine learning techniques, and deep neural networks in particular, are the state-of-the-art for most medical image segmentation tasks. Issues with class imbalance pose a significant challenge in medical datasets, with lesions often occupying a considerably smaller volume relative to the background. Loss functions used in the training of deep learning algorithms differ in their robustness to class imbalance, with direct consequences for model convergence. The most commonly used loss functions for segmentation are based on either the cross entropy loss, Dice loss or a combination of the two. We propose the Unified Focal loss, a new hierarchical framework that generalises Dice and cross entropy-based losses for handling class imbalance. We evaluate our proposed loss function on five publicly available, class imbalanced medical imaging datasets: CVC-ClinicDB, Digital Retinal Images for Vessel Extraction (DRIVE), Breast Ultrasound 2017 (BUS2017), Brain Tumour Segmentation 2020 (BraTS20) and Kidney Tumour Segmentation 2019 (KiTS19). We compare our loss function performance against six Dice or cross entropy-based loss functions, across 2D binary, 3D binary and 3D multiclass segmentation tasks, demonstrating that our proposed loss function is robust to class imbalance and consistently outperforms the other loss functions. Source code is available at: https://github.com/mlyg/unified-focal-loss
Open AccessUnified Focal loss: Generalising Dice and cross entropy-based losses to handle class imbalanced medical image segmentation
Automatic segmentation methods are an important advancement in medical image analysis. Machine learning techniques, and deep neural networks in particular, are the state-of-the-art for most medical image segmentation tasks. Issues with class imbalance pose a significant challenge in medical datasets, with lesions often occupying a considerably smaller volume relative to the background. Loss functions used in the training of deep learning algorithms differ in their robustness to class imbalance, with direct consequences for model convergence. The most commonly used loss functions for segmentation are based on either the cross entropy loss, Dice loss or a combination of the two. We propose the Unified Focal loss, a new hierarchical framework that generalises Dice and cross entropy-based losses for handling class imbalance. We evaluate our proposed loss function on five publicly available, class imbalanced medical imaging datasets: CVC-ClinicDB, Digital Retinal Images for Vessel Extraction (DRIVE), Breast Ultrasound 2017 (BUS2017), Brain Tumour Segmentation 2020 (BraTS20) and Kidney Tumour Segmentation 2019 (KiTS19). We compare our loss function performance against six Dice or cross entropy-based loss functions, across 2D binary, 3D binary and 3D multiclass segmentation tasks, demonstrating that our proposed loss function is robust to class imbalance and consistently outperforms the other loss functions. Source code is available at: https://github.com/mlyg/unified-focal-loss
- Open Access
- 1 October 2021
- Published by Elsevier in Computers in Biology and Medicine
- Vol. 137, 104815
Focus U-Net: A novel dual attention-gated CNN for polyp segmentation during colonoscopy
- 1 October 2021
- Published by Elsevier in Computers in Biology and Medicine
- Vol. 137, 104815
Open AccessFocus U-Net: A novel dual attention-gated CNN for polyp segmentation during colonoscopy