Feature augmentation for imbalanced classification with conditional mixture WGANs

Zhang, Yinghui; Sun, Bo; Xiao, Yunhua; Xiao, Rong; Wei, Yungang

doi:10.1016/j.image.2019.03.010

Cited by 23 publications

(16 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Besides geometric transformations, more complex guided-augmentation methods may be used for DA, such as GAN [23]. In Reference [24], a general framework of DA using GANs in feature space for imbalanced classification is proposed. Experiments are conducted on three databases, i.e., SVHN, FER2013 and Amazon Review of Instant Video, showing the significant improvement with feature augmentation of GANs.…”

Section: Data Augmentation For Fermentioning

confidence: 99%

Evaluation of Data Augmentation Techniques for Facial Expression Recognition Systems

2020

View full text Add to dashboard Cite

Most Facial Expression Recognition (FER) systems rely on machine learning approaches that require large databases for an effective training. As these are not easily available, a good solution is to augment the databases with appropriate data augmentation (DA) techniques, which are typically based on either geometric transformation or oversampling augmentations (e.g., generative adversarial networks (GANs)). However, it is not always easy to understand which DA technique may be more convenient for FER systems because most state-of-the-art experiments use different settings which makes the impact of DA techniques not comparable. To advance in this respect, in this paper, we evaluate and compare the impact of using well-established DA techniques on the emotion recognition accuracy of a FER system based on the well-known VGG16 convolutional neural network (CNN). In particular, we consider both geometric transformations and GAN to increase the amount of training images. We performed cross-database evaluations: training with the "augmented" KDEF database and testing with two different databases (CK+ and ExpW). The best results were obtained combining horizontal reflection, translation and GAN, bringing an accuracy increase of approximately 30%. This outperforms alternative approaches, except for the one technique that could however rely on a quite bigger database.

show abstract

Section: Data Augmentation For Fermentioning

confidence: 99%

Evaluation of Data Augmentation Techniques for Facial Expression Recognition Systems

2020

View full text Add to dashboard Cite

show abstract

“…• DELTA [13] • DIFA+cMWGAN: The adversarial feature augmentation method based on [15]. We modified the structures of GAN with cMWGAN [14].…”

Section: Results and Comparisonsmentioning

confidence: 99%

“…A feature extractor is trained with the source dataset under supervised learning, and then a feature generator for the unlabeled target dataset is trained in the convolutional GAN (CGAN) framework [19] against the feature extractor. Zhang et al [14] developed a more general feature generation framework for imbalanced classification, inspired by the adversarial feature augmentation approach. These methods can generate domain-invariant features without considering the modality of the data.…”

Section: Related Workmentioning

confidence: 99%

“…Unlike data augmentation, adversarial feature augmentation generates domain-invariant features, increasing the size of the minority classes in the feature space without considering the modality of the input data [13]. This augmentation was adopted in supervised learning, especially for the class imbalance problem [14]. Although the feature augmentation method performs better than the unsupervised domain adaptation method by generating domain-invariant and modality-free features, the generated feature maps from the outer fully connected layer do not have sufficient feature dimensions to represent the data distribution with a small amount of minority class data.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Discriminative feature generation for classification of imbalanced data

Suh¹,

Lukowicz²,

Lee³

2020

Preprint

View full text Add to dashboard Cite

The data imbalance problem is a frequent bottleneck in the classification performance of neural networks. In this paper, we propose a novel supervised discriminative feature generation (DFG) method for a minority class dataset. DFG is based on the modified structure of a generative adversarial network consisting of four independent networks: generator, discriminator, feature extractor, and classifier. To augment the selected discriminative features of the minority class data by adopting an attention mechanism, the generator for the class-imbalanced target task is trained, and the feature extractor and classifier are regularized using the pre-trained features from a large source data. The experimental results show that the DFG generator enhances the augmentation of the label-preserved and diverse features, and the classification results are significantly improved on the target task. The feature generation model can contribute greatly to the development of data augmentation methods through discriminative feature generation and supervised attention methods.

show abstract

“…A few pioneering works propose generative-based feature augmentation approaches for domain adaptation [66], imbalanced classification [79], and few-shot learning [8]. Feature normalization plays an important role in neural network training [27,34,49,36].…”

Section: Related Workmentioning

confidence: 99%

Adversarial Learning and Augmentation for Speaker Recognition

Chien¹,

Peng²

2018

The Speaker and Language Recognition Workshop (Odyssey 2018)

View full text Add to dashboard Cite

This paper develops a new generative adversarial network (GAN) to artificially generate i-vectors to deal with the issue of unbalanced or insufficient data in speaker recognition based on the probabilistic linear discriminant analysis (PLDA). Data augmentation is performed to improve system robustness over the variations of i-vectors under different number of training utterances. Our idea is to incorporate the class label into GAN which involves a minimax optimization problem for adversarial learning. We build a generator and a discriminator where the class conditional i-vectors are produced by the generator such that the discriminator can not distinguish them as the fake samples. In particular, multiple learning objectives are optimized to build a specialized deep model for model regularization in speaker recognition. In addition to the minimax optimization of adversarial loss, the posterior probabilities of class labels given real and fake samples are maximized. The cosine similarity between real and fake i-vectors is also minimized to preserve the quality of the generated i-vector. The loss functions for data reconstruction and Gaussian regularization in PLDA model are minimized. The experiments illustrate the merit of multi-objective learning for deep adversarial augmentation for speaker recognition.

show abstract

Feature augmentation for imbalanced classification with conditional mixture WGANs

Cited by 23 publications

References 16 publications

Evaluation of Data Augmentation Techniques for Facial Expression Recognition Systems

Evaluation of Data Augmentation Techniques for Facial Expression Recognition Systems

Discriminative feature generation for classification of imbalanced data

Adversarial Learning and Augmentation for Speaker Recognition

Contact Info

Product

Resources

About