1
|
Fu J, Peng H, Li B, Liu Z, Lugu R, Wang J, Ramírez-de-Arellano A. Multitask Adversarial Networks Based on Extensive Nonlinear Spiking Neuron Models. Int J Neural Syst 2024; 34:2450032. [PMID: 38624267 DOI: 10.1142/s0129065724500321] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 04/17/2024]
Abstract
Deep learning technology has been successfully used in Chest X-ray (CXR) images of COVID-19 patients. However, due to the characteristics of COVID-19 pneumonia and X-ray imaging, the deep learning methods still face many challenges, such as lower imaging quality, fewer training samples, complex radiological features and irregular shapes. To address these challenges, this study first introduces an extensive NSNP-like neuron model, and then proposes a multitask adversarial network architecture based on ENSNP-like neurons for chest X-ray images of COVID-19, called MAE-Net. The MAE-Net serves two tasks: (i) converting low-quality CXR images to high-quality images; (ii) classifying CXR images of COVID-19. The adversarial architecture of MAE-Net uses two generators and two discriminators, and two new loss functions have been introduced to guide the optimization of the network. The MAE-Net is tested on four benchmark COVID-19 CXR image datasets and compared them with eight deep learning models. The experimental results show that the proposed MAE-Net can enhance the conversion quality and the accuracy of image classification results.
Collapse
Affiliation(s)
- Jun Fu
- School of Computer and Software Engineering, Xihua University, Chengdu 610039, P. R. China
| | - Hong Peng
- School of Computer and Software Engineering, Xihua University, Chengdu 610039, P. R. China
| | - Bing Li
- School of Computer and Software Engineering, Xihua University, Chengdu 610039, P. R. China
| | - Zhicai Liu
- School of Computer and Software Engineering, Xihua University, Chengdu 610039, P. R. China
| | - Rikong Lugu
- School of Computer and Software Engineering, Xihua University, Chengdu 610039, P. R. China
| | - Jun Wang
- School of Electrical Engineering and Electronic Information, Xihua University, Chengdu 610039, P. R. China
| | - Antonio Ramírez-de-Arellano
- Research Group of Natural Computing, Department of Computer Science and Artificial Intelligence, University of Seville, Sevilla 41012, Spain
| |
Collapse
|
2
|
Hetz MJ, Bucher TC, Brinker TJ. Multi-domain stain normalization for digital pathology: A cycle-consistent adversarial network for whole slide images. Med Image Anal 2024; 94:103149. [PMID: 38574542 DOI: 10.1016/j.media.2024.103149] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/21/2022] [Revised: 12/11/2023] [Accepted: 03/20/2024] [Indexed: 04/06/2024]
Abstract
The variation in histologic staining between different medical centers is one of the most profound challenges in the field of computer-aided diagnosis. The appearance disparity of pathological whole slide images causes algorithms to become less reliable, which in turn impedes the wide-spread applicability of downstream tasks like cancer diagnosis. Furthermore, different stainings lead to biases in the training which in case of domain shifts negatively affect the test performance. Therefore, in this paper we propose MultiStain-CycleGAN, a multi-domain approach to stain normalization based on CycleGAN. Our modifications to CycleGAN allow us to normalize images of different origins without retraining or using different models. We perform an extensive evaluation of our method using various metrics and compare it to commonly used methods that are multi-domain capable. First, we evaluate how well our method fools a domain classifier that tries to assign a medical center to an image. Then, we test our normalization on the tumor classification performance of a downstream classifier. Furthermore, we evaluate the image quality of the normalized images using the Structural similarity index and the ability to reduce the domain shift using the Fréchet inception distance. We show that our method proves to be multi-domain capable, provides a very high image quality among the compared methods, and can most reliably fool the domain classifier while keeping the tumor classifier performance high. By reducing the domain influence, biases in the data can be removed on the one hand and the origin of the whole slide image can be disguised on the other, thus enhancing patient data privacy.
Collapse
Affiliation(s)
- Martin J Hetz
- Division of Digital Biomarkers for Oncology, German Cancer Research Center (DKFZ), Heidelberg, Germany
| | - Tabea-Clara Bucher
- Division of Digital Biomarkers for Oncology, German Cancer Research Center (DKFZ), Heidelberg, Germany
| | - Titus J Brinker
- Division of Digital Biomarkers for Oncology, German Cancer Research Center (DKFZ), Heidelberg, Germany.
| |
Collapse
|
3
|
Berijanian M, Schaadt NS, Huang B, Lotz J, Feuerhake F, Merhof D. Unsupervised many-to-many stain translation for histological image augmentation to improve classification accuracy. J Pathol Inform 2023; 14:100195. [PMID: 36844704 PMCID: PMC9947329 DOI: 10.1016/j.jpi.2023.100195] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/15/2022] [Revised: 01/19/2023] [Accepted: 01/20/2023] [Indexed: 01/26/2023] Open
Abstract
Background Deep learning tasks, which require large numbers of images, are widely applied in digital pathology. This poses challenges especially for supervised tasks since manual image annotation is an expensive and laborious process. This situation deteriorates even more in the case of a large variability of images. Coping with this problem requires methods such as image augmentation and synthetic image generation. In this regard, unsupervised stain translation via GANs has gained much attention recently, but a separate network must be trained for each pair of source and target domains. This work enables unsupervised many-to-many translation of histopathological stains with a single network while seeking to maintain the shape and structure of the tissues. Methods StarGAN-v2 is adapted for unsupervised many-to-many stain translation of histopathology images of breast tissues. An edge detector is incorporated to motivate the network to maintain the shape and structure of the tissues and to have an edge-preserving translation. Additionally, a subjective test is conducted on medical and technical experts in the field of digital pathology to evaluate the quality of generated images and to verify that they are indistinguishable from real images. As a proof of concept, breast cancer classifiers are trained with and without the generated images to quantify the effect of image augmentation using the synthetized images on classification accuracy. Results The results show that adding an edge detector helps to improve the quality of translated images and to preserve the general structure of tissues. Quality control and subjective tests on our medical and technical experts show that the real and artificial images cannot be distinguished, thereby confirming that the synthetic images are technically plausible. Moreover, this research shows that, by augmenting the training dataset with the outputs of the proposed stain translation method, the accuracy of breast cancer classifier with ResNet-50 and VGG-16 improves by 8.0% and 9.3%, respectively. Conclusions This research indicates that a translation from an arbitrary source stain to other stains can be performed effectively within the proposed framework. The generated images are realistic and could be employed to train deep neural networks to improve their performance and cope with the problem of insufficient numbers of annotated images.
Collapse
Affiliation(s)
- Maryam Berijanian
- Department of Computational Mathematics, Science and Engineering (CMSE), Michigan State University, East Lansing, USA,Institute of Imaging and Computer Vision, RWTH Aachen University, Aachen, Germany
| | | | - Boqiang Huang
- Institute of Image Analysis and Computer Vision, Faculty of Informatics and Data Science, University of Regensburg, Regensburg, Germany
| | - Johannes Lotz
- Fraunhofer Institute for Digital Medicine MEVIS, Lübeck, Germany
| | - Friedrich Feuerhake
- Institute for Pathology, Hannover Medical School, Hannover, Germany,Institute for Neuropathology, University Clinic Freiburg, Freiburg, Germany
| | - Dorit Merhof
- Institute of Image Analysis and Computer Vision, Faculty of Informatics and Data Science, University of Regensburg, Regensburg, Germany,Fraunhofer Institute for Digital Medicine MEVIS, Bremen, Germany,Corresponding author at: University of Regensburg, 93040 Regensburg, Germany.
| |
Collapse
|
4
|
de Bel T, Bokhorst JM, van der Laak J, Litjens G. Residual cyclegan for robust domain transformation of histopathological tissue slides. Med Image Anal 2021; 70:102004. [PMID: 33647784 DOI: 10.1016/j.media.2021.102004] [Citation(s) in RCA: 24] [Impact Index Per Article: 8.0] [Reference Citation Analysis] [What about the content of this article? (0)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/15/2020] [Revised: 02/10/2021] [Accepted: 02/15/2021] [Indexed: 12/26/2022]
Abstract
Variation between stains in histopathology is commonplace across different medical centers. This can have a significant effect on the reliability of machine learning algorithms. In this paper, we propose to reduce performance variability by using -consistent generative adversarial (CycleGAN) networks to remove staining variation. We improve upon the regular CycleGAN by incorporating residual learning. We comprehensively evaluate the performance of our stain transformation method and compare its usefulness in addition to extensive data augmentation to enhance the robustness of tissue segmentation algorithms. Our steps are as follows: first, we train a model to perform segmentation on tissue slides from a single source center, while heavily applying augmentations to increase robustness to unseen data. Second, we evaluate and compare the segmentation performance on data from other centers, both with and without applying our CycleGAN stain transformation. We compare segmentation performances in a colon tissue segmentation and kidney tissue segmentation task, covering data from 6 different centers. We show that our transformation method improves the overall Dice coefficient by 9% over the non-normalized target data and by 4% over traditional stain transformation in our colon tissue segmentation task. For kidney segmentation, our residual CycleGAN increases performance by 10% over no transformation and around 2% compared to the non-residual CycleGAN.
Collapse
|
5
|
Abstract
BACKGROUND Automatic extraction of biomedical events from literature, that allows for faster update of the latest discoveries automatically, is a heated research topic now. Trigger word recognition is a critical step in the process of event extraction. Its performance directly influences the results of the event extraction. In general, machine learning-based trigger recognition approaches such as neural networks must to be trained on a dataset with plentiful annotations to achieve high performances. However, the problem of the datasets in wide coverage event domains is that their annotations are insufficient and imbalance. One of the methods widely used to deal with this problem is transfer learning. In this work, we aim to extend the transfer learning to utilize multiple source domains. Multiple source domain datasets can be jointly trained to help achieve a higher recognition performance on a target domain with wide coverage events. RESULTS Based on the study of previous work, we propose an improved multi-source domain neural network transfer learning architecture and a training approach for biomedical trigger detection task, which can share knowledge between the multi-source and target domains more comprehensively. We extend the ability of traditional adversarial networks to extract common features between source and target domains, when there is more than one dataset in the source domains. Multiple feature extraction channels to simultaneously capture global and local common features are designed. Moreover, under the constraint of an extra classifier, the multiple local common feature sub-channels can extract and transfer more diverse common features from the related multi-source domains effectively. In the experiments, MLEE corpus is used to train and test the proposed model to recognize the wide coverage triggers as a target dataset. Other four corpora with the varying degrees of relevance with MLEE from different domains are used as source datasets, respectively. Our proposed approach achieves recognition improvement compared with traditional adversarial networks. Moreover, its performance is competitive compared with the results of other leading systems on the same MLEE corpus. CONCLUSIONS The proposed Multi-Source Transfer Learning-based Trigger Recognizer (MSTLTR) can further improve the performance compared with the traditional method, when the source domains are more than one. The most essential improvement is that our approach represents common features in two aspects: the global common features and the local common features. Hence, these more sharable features improve the performance and generalization of the model on the target domain effectively.
Collapse
Affiliation(s)
- Yifei Chen
- School of Information Engineering, Nanjing Audit University, 86 West Yushan Road, Nanjing, China.
| |
Collapse
|
6
|
Liu F, Xu M, Li G, Pei J, Shi L, Zhao R. Adversarial symmetric GANs: Bridging adversarial samples and adversarial networks. Neural Netw 2020; 133:148-156. [PMID: 33217683 DOI: 10.1016/j.neunet.2020.10.016] [Citation(s) in RCA: 4] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/31/2019] [Revised: 09/19/2020] [Accepted: 10/29/2020] [Indexed: 11/29/2022]
Abstract
Generative adversarial networks have achieved remarkable performance on various tasks but suffer from training instability. Despite many training strategies proposed to improve training stability, this issue remains as a challenge. In this paper, we investigate the training instability from the perspective of adversarial samples and reveal that adversarial training on fake samples is implemented in vanilla GANs, but adversarial training on real samples has long been overlooked. Consequently, the discriminator is extremely vulnerable to adversarial perturbation and the gradient given by the discriminator contains non-informative adversarial noises, which hinders the generator from catching the pattern of real samples. Here, we develop adversarial symmetric GANs (AS-GANs) that incorporate adversarial training of the discriminator on real samples into vanilla GANs, making adversarial training symmetrical. The discriminator is therefore more robust and provides more informative gradient with less adversarial noise, thereby stabilizing training and accelerating convergence. The effectiveness of the AS-GANs is verified on image generation on CIFAR-10, CIFAR-100, CelebA, and LSUN with varied network architectures. Not only the training is more stabilized, but the FID scores of generated samples are consistently improved by a large margin compared to the baseline. Theoretical analysis is also conducted to explain why AS-GAN can improve training. The bridging of adversarial samples and adversarial networks provides a new approach to further develop adversarial networks.
Collapse
Affiliation(s)
- Faqiang Liu
- Department of Precision Instrument, Tsinghua University, Beijing, 100084, China; Center for Brain Inspired Computing Research, Tsinghua University, Beijing, 100084, China; Beijing Innovation Center for Future Chip, Beijing, 100084, China
| | - Mingkun Xu
- Department of Precision Instrument, Tsinghua University, Beijing, 100084, China; Center for Brain Inspired Computing Research, Tsinghua University, Beijing, 100084, China; Beijing Innovation Center for Future Chip, Beijing, 100084, China
| | - Guoqi Li
- Department of Precision Instrument, Tsinghua University, Beijing, 100084, China; Center for Brain Inspired Computing Research, Tsinghua University, Beijing, 100084, China; Beijing Innovation Center for Future Chip, Beijing, 100084, China
| | - Jing Pei
- Department of Precision Instrument, Tsinghua University, Beijing, 100084, China; Center for Brain Inspired Computing Research, Tsinghua University, Beijing, 100084, China; Beijing Innovation Center for Future Chip, Beijing, 100084, China
| | - Luping Shi
- Department of Precision Instrument, Tsinghua University, Beijing, 100084, China; Center for Brain Inspired Computing Research, Tsinghua University, Beijing, 100084, China; Beijing Innovation Center for Future Chip, Beijing, 100084, China
| | - Rong Zhao
- Department of Precision Instrument, Tsinghua University, Beijing, 100084, China; Center for Brain Inspired Computing Research, Tsinghua University, Beijing, 100084, China; Beijing Innovation Center for Future Chip, Beijing, 100084, China.
| |
Collapse
|