Recurrent generative adversarial network for learning imbalanced medical image semantic segmentation

Rezaei, Mina; Yang, Haojin; Meinel, Christoph

doi:10.1007/s11042-019-7305-1

Cited by 59 publications

(38 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Similar to this work, M. Rezaei et al [225] used similar loss function by mixing adversarial loss and weighted categorical accuracy loss to handle imbalanced training dataset of whole heart segmentation task. Balancing through ensemble learning by combining two discriminator to improve their generalization ability of the GAN was tested by M. Rezaei et al [226] in medical image semantic segmentation task.…”

Section: Pixel-wise Class Imbalancementioning

confidence: 99%

“…Few minority-many majority class imbalance Cycle GAN [194] Emotion classification Few minority-many majority class imbalance DCGAN [195] Weather classification Few minority-many majority class imbalance DCGAN + Ensemble learning [196] Weather classification Few minority-many majority class imbalance DCGAN [190] Chest pathology classification Few minority-many majority class imbalance DCGAN [197] liver lesion classification Many majority-Few minority class imbalance DCGAN [198] Skin [225] Heart image segmentation Imbalance due to occlusions SeGAN [229] Invisible part generation and Segmentation Imbalance due to occlusions…”

Section: Multi Class Classificationmentioning

confidence: 99%

See 1 more Smart Citation

A Survey on Generative Adversarial Networks for imbalance problems in computer vision tasks

Sampath

Maurtua

Martín

et al. 2020

Preprint

View full text Add to dashboard Cite

Any computer vision application development starts off by acquiring images and data, then preprocessing and pattern recognition steps to perform a task. When the acquired images are highly imbalanced and not adequate, the desired task may not be achievable. Unfortunately, the occurrence of imbalance problems in acquired image datasets in certain complex real-world problems such as anomaly detection, emotion recognition, medical image analysis, fraud detection, metallic surface defect detection, disaster prediction, etc., are inevitable. The performance of computer vision algorithms can significantly deteriorate when the training dataset is imbalanced. In recent years, Generative Adversarial Networks (GANs) have gained immense attention by researchers across a variety of application domains due to their capability to model complex real-world image data. It is particularly important that GANs can not only be used to generate synthetic images, but also its fascinating adversarial learning idea showed good potential in restoring balance in imbalanced datasets. In this paper, we examine the most recent developments of GANs based techniques for addressing imbalance problems in image data. The real-world challenges and implementations of synthetic image generation based on GANs are extensively covered in this survey. Our survey first introduces various imbalance problems in computer vision tasks and its existing solutions, and then examine key concepts such as deep generative image models and GANs. After that, we propose a taxonomy to summarize GANs based techniques for addressing imbalance problems in computer vision tasks into three major categories: 1. Image level imbalances in classification, 2. object level imbalances in object detection and 3. pixel level imbalances in segmentation tasks. We elaborate the imbalance problems of each group, and further provide GANs based solutions in each group. Readers will understand how GANs based techniques can handle the problem of imbalances and boost performance of the computer vision algorithms.

show abstract

Section: Pixel-wise Class Imbalancementioning

confidence: 99%

Section: Multi Class Classificationmentioning

confidence: 99%

A Survey on Generative Adversarial Networks for imbalance problems in computer vision tasks

Sampath

Maurtua

Martín

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…Algorithmic methods have included cost-sensitive learning and ensemble learning. The cost-sensitive learning is typically used with accuracy loss [ 2 ], Dice coefficient loss [ 3 ], and asymmetric similarity loss [ 4 ] to modify the distribution of the training data based on a mis-classification cost. However, in the case of image segmentation, losses such as mean surface distance or Hausdorff surface distance are more appropriate.…”

Section: Related Workmentioning

confidence: 99%

Generative multi-adversarial network for striking the right balance in abdominal image segmentation

et al. 2020

Self Cite

View full text Add to dashboard Cite

Purpose The identification of abnormalities that are relatively rare within otherwise normal anatomy is a major challenge for deep learning in the semantic segmentation of medical images. The small number of samples of the minority classes in the training data makes the learning of optimal classification challenging, while the more frequently occurring samples of the majority class hamper the generalization of the classification boundary between infrequently occurring target objects and classes. In this paper, we developed a novel generative multi-adversarial network, called Ensemble-GAN, for mitigating this class imbalance problem in the semantic segmentation of abdominal images. Method The Ensemble-GAN framework is composed of a single-generator and a multi-discriminator variant for handling the class imbalance problem to provide a better generalization than existing approaches. The ensemble model aggregates the estimates of multiple models by training from different initializations and losses from various subsets of the training data. The single generator network analyzes the input image as a condition to predict a corresponding semantic segmentation image by use of feedback from the ensemble of discriminator networks. To evaluate the framework, we trained our framework on two public datasets, with different imbalance ratios and imaging modalities: the Chaos 2019 and the LiTS 2017. Result In terms of the F1 score, the accuracies of the semantic segmentation of healthy spleen, liver, and left and right kidneys were 0.93, 0.96, 0.90 and 0.94, respectively. The overall F1 scores for simultaneous segmentation of the lesions and liver were 0.83 and 0.94, respectively. Conclusion The proposed Ensemble-GAN framework demonstrated outstanding performance in the semantic segmentation of medical images in comparison with other approaches on popular abdominal imaging benchmarks. The Ensemble-GAN has the potential to segment abdominal images more accurately than human experts.

show abstract

“…Similar to this work, M. Rezaei [222] used similar loss function by mixing adversarial loss and weighted categorical accuracy loss to handle imbalanced training dataset of whole heart segmentation task. Balancing through ensemble learning by combining two discriminator to improve their generalization ability of the GAN was tested by M. Rezaei et al [223] in medical image semantic segmentation task.…”

Section: Pixel Level Imbalances In Segmentationmentioning

confidence: 99%

“…Few minority-many majority class imbalance DCGAN [187] Chest pathology classification Few minority-many majority class imbalance DCGAN [194] liver lesion classification Many majority-Few minority class imbalance DCGAN [195] Skin lesion classification Many majority-Many minority class imbalance Cycle-GAN [186] Plant disease classification Many majority-Many minority class imbalance WGAN-GP [197] Multi class classification Many majority-Many minority class imbalance CE-GAN [200] Multi class classification [222] Heart image segmentation Imbalance due to occlusions SeGAN [226] Invisible part generation and Segmentation Imbalance due to occlusions…”

Section: Multi Class Classificationmentioning

confidence: 99%

A Survey on Generative Adversarial Networks for Imbalance Problems in Computer Vision Tasks

Sampath

Maurtua

Martín

et al. 2020

Preprint

View full text Add to dashboard Cite

Any computer vision application development starts off by acquiring images and data, then preprocessingand pattern recognition steps to perform a task. When the acquired image is highly imbalanced and notadequate, the desired task may not be achievable. Unfortunately, the occurrence of imbalance problems inacquired image datasets in certain complex real-world problems such as anomaly detection, emotionrecognition, medical image analysis, fraud detection, metallic surface defect detection, disaster prediction,etc., are inevitable. The performance of computer vision algorithms can significantly deteriorate when thetraining dataset is imbalanced. In recent years, Generative Adversarial Networks (GANs) have gainedimmense attention by researchers across a variety of application domains due to their capability to modelcomplex real-world image data. It is particularly important that GANs can not only be used to generatesynthetic images, but also its fascinating adversarial learning idea showed good potential in restoringbalance in imbalanced datasets.In this paper, we examine the most recent developments of GANs based techniques for addressingimbalance problems in image data. The real-world challenges and implementations of synthetic imagegeneration based on GANs are extensively covered in this survey. Our survey first introduces variousimbalance problems in computer vision tasks and its existing solutions, and then examine key conceptssuch as deep generative image models and GANs. After that, we propose taxonomy to summarize GANsbased techniques for addressing imbalance problems in computer vision tasks into three major categories:Image level imbalances in classification, object level imbalances in object detection and pixel levelimbalances in segmentation tasks. We elaborate the imbalance problems of each group, and furtherprovide GANs based solutions in each group. Readers will understand how GANs based techniques canhandle the problem of imbalances and boost performance of the computer vision algorithms.

show abstract

Recurrent generative adversarial network for learning imbalanced medical image semantic segmentation

Cited by 59 publications

References 37 publications

A Survey on Generative Adversarial Networks for imbalance problems in computer vision tasks

A Survey on Generative Adversarial Networks for imbalance problems in computer vision tasks

Generative multi-adversarial network for striking the right balance in abdominal image segmentation

A Survey on Generative Adversarial Networks for Imbalance Problems in Computer Vision Tasks

Contact Info

Product

Resources

About