The DOI matches its Crossref record; 77 of the 83 references it cites were found in a registry or an index.
A systematic study of the class imbalance problem in convolutional neural networks
Open accessAbstract
In this study, we systematically investigate the impact of class imbalance on classification performance of convolutional neural networks (CNNs) and compare frequently used methods to address the issue. Class imbalance is a common problem that has been comprehensively studied in classical machine learning, yet very limited systematic research is available in the context of deep learning. In our study, we use three benchmark datasets of increasing complexity, MNIST, CIFAR-10 and ImageNet, to investigate the effects of imbalance on classification and perform an extensive comparison of several methods to address the issue: oversampling, undersampling, two-phase training, and thresholding that compensates for prior class probabilities. Our main evaluation metric is area under the receiver operating characteristic curve (ROC AUC) adjusted to multi-class tasks since overall accuracy metric is associated with notable difficulties in the context of imbalanced data. Based on results from our experiments we conclude that (i) the effect of class imbalance on classification performance is detrimental; (ii) the method of addressing class imbalance that emerged as dominant in almost all analyzed scenarios was oversampling; (iii) oversampling should be applied to the level that completely eliminates the imbalance, whereas the optimal undersampling ratio depends on the extent of imbalance; (iv) as opposed to some classical machine learning models, oversampling does not cause overfitting of CNNs; (v) thresholding should be applied to compensate for prior class probabilities when overall number of properly classified cases is of interest.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“A memory of the past is needed to rectify past classes' scores.”
From A Comprehensive Study of Class Incremental Learning Algorithms for Visual Tasks (2020) · cites this paper as [12]
In the citing paper
“Here, we introduced a competitive method where classification bias in favor of new classes is reduced by using prior class probabilities [12].”
From A Comprehensive Study of Class Incremental Learning Algorithms for Visual Tasks (2020) · cites this paper as [12]
In the citing paper
“Finally, the highly-discussed dimension target class balance pays tribute to the technical requirements of ML140,141,144,150,159.”
From The METRIC-framework for assessing data quality for trustworthy AI in medicine: A systematic review (2024) · cites this paper as [141]
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 1×
“And for all the cases, the classes chosen to be minority were the ones with the lowest label value.”
BackgroundVery Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 114,148 · cited 1×
“A standard CNN architecture comprises fully connected layers and a number of blocks consisting of convolution kernels, activation function layer and max pooling [22, 23, 24].”
MethodImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155
Scikit-learn: Machine Learning in Python
Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Müller, A., Nothman, J., Louppe, G., et al. (2012) · arXiv · doi:10.48550/arxiv.1201.0490 · Cited by 71,016 · cited 1×
“We used a specific implementation of the ROC AUC available in scikit-learn python package [65].”
MethodGradient-based learning applied to document recognition
Lecun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998) · Proceedings of the IEEE · doi:10.1109/5.726791 · Cited by 58,805 · cited 2×
“Images The CNN model that we use for MNIST is the modern version of LeNet-5 [52].”
MethodImageNet Large Scale Visual Recognition Challenge
Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., Huang, Z., Karpathy, A., Khosla, A., Bernstein, M., et al. (2015) · International Journal of Computer Vision · doi:10.1007/s11263-015-0816-y · Cited by 39,380 · cited 1×
“In our study, we used three benchmark datasets: MNIST [52], CIFAR-10 [53] and ImageNet Large Scale Visual Recognition Challenge (ILSVRC) 2012 [54].”
MethodSMOTE synthetic minority over-sampling technique
Chawla, N. V., Bowyer, K. W., Hall, L. O., & Kegelmeyer, W. P. (2002) · Journal of Artificial Intelligence Research · doi:10.1613/jair.953 · Cited by 32,865 · cited 3×
“For classical machine learning models it was shown that oversampling can cause overfitting, especially for minority classes [33].”
BackgroundMethodDelving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification
He, K., Zhang, X., Ren, S., & Sun, J. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.123 · Cited by 19,741 · cited 1×
“Network weights were initialized with Kaiming (also known as MSRA) initialization procedure [63].”
MethodVisualizing and understanding convolutional networks
Zeiler, M. D., & Fergus, R. (2014) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-10590-1_53 · Cited by 15,839 · cited 1×
“This property allows CNNs to learn hierarchical representations [2].”
BackgroundBatch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate ShiftPreprint
Ioffe, S. (2024) · arXiv (Cornell University) · doi:10.57702/o9raffed · Cited by 15,273
(2010) · Cited by 14,134
Proceedings of IEEE International Conference on Computer Vision
(1995) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.1995.466933 · Cited by 13,117
Advances in Neural Information Processing Systems 25
Zhang, Y., Sutton, C. A., Storkey, A., & Zoubin Ghahramani, A. (2012) · Cited by 12,653
Backpropagation Applied to Handwritten Zip Code Recognition
LeCun, Y., Boser, B., Denker, J. S., Henderson, D., Howard, R. E., Hubbard, W., & Jackel, L. D. (1989) · Neural Computation · doi:10.1162/neco.1989.1.4.541 · Cited by 11,891 · cited 1×
“The standard CNN is built with fully connected layers and a number of blocks consisting of convolutions, activation function layer and max pooling [3, 4, 5].”
MethodCaffe: Convolutional Architecture for Fast Feature Embedding
Jia, Y., Shelhamer, E., Donahue, J., Karayev, S., Long, J., Girshick, R., Guadarrama, S., & Darrell, T. (2014) · Proceedings of the 22nd ACM international conference on Multimedia · doi:10.1145/2647868.2654889 · Cited by 11,077 · cited 1×
“All experiments were implemented in the deep learning framework Caffe [55].”
MethodHe, H., & Garcia, E. A. (2009) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2008.239 · Cited by 10,359 · cited 1×
“Methods for addressing class imbalance can be divided into two main categories [29].”
BackgroundThe use of the area under the ROC curve in the evaluation of machine learning algorithms
Bradley, A. P. (1997) · Pattern Recognition · doi:10.1016/s0031-3203(96)00142-2 · Cited by 7,217
Recent advances in convolutional neural networks
Gu, J., Wang, Z., Kuen, J., Ma, L., Shahroudy, A., Shuai, B., Liu, T., Wang, X., Wang, G., Cai, J., et al. (2018) · Pattern Recognition · doi:10.1016/j.patcog.2017.10.013 · Cited by 6,576
2009 IEEE Conference on Computer Vision and Pattern Recognition
(2009) · doi:10.1109/cvpr14067.2009 · Cited by 5,646
Caffe: Convolutional Architecture for Fast Feature EmbeddingPreprint
Jia, Y., Shelhamer, E., Donahue, J., Karayev, S., Long, J., Girshick, R., Guadarrama, S., & Darrell, T. (2014) · arXiv (Cornell University) · doi:10.48550/arxiv.1408.5093 · Cited by 5,299 · cited 1×
“All experiments were implemented in the deep learning framework Caffe [55].”
MethodBorderline-SMOTE : A New Over-Sampling Method in Imbalanced Data Sets Learning
Han, H., Wang, W.-Y., & Mao, B.-H. (2005) · Lecture notes in computer science · doi:10.1007/11538059_91 · Cited by 4,331 · cited 1×
“Some extensions of this technique were proposed, for example focusing only on examples near the boundary between classes [35].”
MethodStriving for Simplicity: The All Convolutional Net
Springenberg, J. T., Dosovitskiy, A., Brox, T., & Riedmiller, M. (2014) · arXiv · doi:10.48550/arxiv.1412.6806 · Cited by 3,447 · cited 2×
“For CIFAR-10 experiments we use one of the best performing type of CNN model on this dataset, i.e. All-CNN [59].”
MethodThe class imbalance problem: A systematic study1
Japkowicz, N., & Stephen, S. (2002) · Intelligent Data Analysis · doi:10.3233/ida-2002-6504 · Cited by 3,295 · cited 3×
“Na¨ıve version, called random majority undersampling, simply removes a random portion of examples from majority classes [17].”
BackgroundSUN database: Large-scale scene recognition from abbey to zoo
Xiao, J., Hays, J., Ehinger, K. A., Oliva, A., & Torralba, A. (2010) · 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2010.5539970 · Cited by 3,283 · cited 1×
“There are plenty of examples in domains like computer vision [6, 7, 8, 9, 10], medical diagnosis [11, 12], fraud detection [13] and others [14, 15, 16] where this issue is highly significant and the frequency of one class (e.g., cancer) can be 1000 times less than another class (e.g., healthy…”
BackgroundBrain tumor segmentation with Deep Neural Networks
Havaei, M., Davy, A., Warde-Farley, D., Biard, A., Courville, A., Bengio, Y., Pal, C., Jodoin, P.-M., & Larochelle, H. (2017) · Medical Image Analysis · doi:10.1016/j.media.2016.05.004 · Cited by 3,278 · cited 2×
“Recently introduced and successfully applied to CNN training for brain tumor segmentation, is two-phase training [28].”
BackgroundMethodExploratory Undersampling for Class-Imbalance Learning
Liu, X., Wu, J., & Zhou, Z. (2008) · IEEE Transactions on Systems Man and Cybernetics Part B (Cybernetics) · doi:10.1109/tsmcb.2008.2007853 · Cited by 2,525 · cited 1×
“EasyEnsemble and BalanceCascade are methods that train a committee of classifiers on undersampled subsets [50].”
Learning from class-imbalanced data: Review of methods and applications
Haixiang, G., Li, Y., Shang, J. S., Gu, M., Huang, Y., & Bing, G. (2016) · Expert Systems with Applications · doi:10.1016/j.eswa.2016.12.035 · Cited by 2,508 · cited 5×
“In the previous decade the number of publications on imbalance learning problem started growing rapidly [11, 16].”
BackgroundMethodResultInfluential2013 IEEE Conference on Computer Vision and Pattern Recognition
(2013) · doi:10.1109/cvpr19808.2013 · Cited by 2,439
On the momentum term in gradient descent learning algorithms
Ning Qian, A. (1999) · Neural Networks · doi:10.1016/s0893-6080(98)00116-6 · Cited by 2,318 · cited 1×
“9 [56].”
Addressing the Curse of Imbalanced Training Sets: One-Sided Selection.
Kubát, M., & Matwin, S. (1997) · Cited by 2,230
The foundations of cost-sensitive learning
Elkan, C. (2001) · Cited by 1,945
SMOTEBoost: Improving Prediction of the Minority Class in Boosting
Chawla, N. V., Lazarević, A. M., Hall, L., & Bowyer, K. W. (2003) · Lecture notes in computer science · doi:10.1007/978-3-540-39804-2_12 · Cited by 1,763 · cited 1×
“SMOTEBoost, on the other hand, is a combination of boosting and SMOTE oversampling [51].”
BackgroundData Mining and Knowledge Discovery Handbook
Maimon, O., & Rokach, L. (2005) · Springer eBooks · doi:10.1007/b107408 · Cited by 1,559
2011 International Joint Conference on Neural Networks
Boubacar, H. A., Lecoeuche, S., & Maouche, S. (2010) · IEEE Transactions on Neural Networks · doi:10.1109/tnn.2010.2063350 · Cited by 1,481
Deep learning for digital pathology image analysis: A comprehensive tutorial with selected use cases
Janowczyk, A. R., & Madabhushi, A. (2016) · Journal of Pathology Informatics · doi:10.4103/2153-3539.186902 · Cited by 1,387 · cited 1×
“One of the most commonly used method in deep learning [16, 30, 31, 32].”
MethodTraining cost-sensitive neural networks with methods addressing the class imbalance problem
Zhou, Z., & Liu, X. (2006) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2006.17 · Cited by 1,337
Machine Learning for the Detection of Oil Spills in Satellite Radar Images
Kubát, M., Holte, R. C., & Matwin, S. (1998) · Machine Learning · doi:10.1023/a:1007452223027 · Cited by 1,266
Data Mining for Imbalanced Datasets: An Overview
Chawla, N. V. (2009) · Springer eBooks · doi:10.1007/978-0-387-09823-4_45 · Cited by 1,197 · cited 2×
“However, it has some significant and long acknowledged limitations, particularly in the context of imbalanced datasets [19].”
BackgroundAge and gender classification using convolutional neural networks
Levi, G., & Hassncer, T. (2015) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw.2015.7301352 · Cited by 1,194 · cited 1×
“The most commonly used method in both classical machine learning and deep learning is oversampling [11, 35, 36, 37].”
MethodNeural Network Classifiers Estimate Bayesian a posteriori Probabilities
Richard, M. D., & Lippmann, R. P. (1991) · Neural Computation · doi:10.1162/neco.1991.3.4.461 · Cited by 1,004 · cited 2×
“However, the most basic version simply compensates for prior class probabilities [43].”
BackgroundCost-Sensitive Learning of Deep Feature Representations From Imbalanced Data
Khan, S. H., Hayat, M., Bennamoun, M., Sohel, F. A., & Togneri, R. B. (2017) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2017.2732482 · Cited by 996 · cited 1×
“Some previous studies showed results on cost sensitive learning of deep neural networks [24, 25, 26].”
ResultGoodfellow, I. J., Warde-Farley, D., Mirza, M., Courville, A., & Bengio, Y. (2013) · arXiv · doi:10.48550/arxiv.1302.4389 · Cited by 935 · cited 1×
“There are exactly 5000 training and 1000 test examples in each class. We do not use any data augmentation but follow standard preprocessing comprising global contrast normalization and ZCA whitening [58]. For CIFAR-10 experiments we use one of the best performing type of CNN model on this dataset, i.e. All-CNN [59]. T…”
MethodNeural Networks: Tricks of the Trade
Orr, G., & Müller, K.-R. (1998) · Lecture notes in computer science · doi:10.1007/3-540-49430-8 · Cited by 817
C4.5, Class Imbalance, and Cost Sensitivity: Why Under-Sampling beats Over-Sampling
Drummond, C., & Holte, R. C. (2003) · Cited by 800
Mazurowski, M. A., Habas, P. A., Zurada, J. M., Lo, J. Y.-C., Baker, J. A., & Tourassi, G. D. (2007) · Neural Networks · doi:10.1016/j.neunet.2007.12.031 · Cited by 797 · cited 3×
“It has been established that class imbalance can have significant detrimental effect on training of traditional classifiers [17] including multi-layer perceptrons [18].”
BackgroundMethodSelf-citationProceedings of the 22nd ACM international conference on Multimedia
(2014) · doi:10.1145/2647868 · Cited by 741
Class imbalances versus small disjuncts
Jo, T., & Japkowicz, N. (2004) · ACM SIGKDD Explorations Newsletter · doi:10.1145/1007730.1007737 · Cited by 687 · cited 1×
“Cluster-based oversampling first clusters the dataset and then oversamples each cluster separately [36].”
BackgroundData mining for direct marketing: problems and solutions
Ling, C. X., & Cheng-hui, L. (1998) · Cited by 661
Learning from imbalanced data sets with boosting and data generation
Guo, H., & Viktor, H. L. (2004) · ACM SIGKDD Explorations Newsletter · doi:10.1145/1007730.1007736 · Cited by 609 · cited 1×
“DataBoost-IM, on the other hand, identifies difficult examples with boosting preprocessing and uses them to generate synthetic data [37].”
MethodTree Induction for Probability-Based Ranking
Provost, F., & Domingos, P. (2003) · Machine Learning · doi:10.1023/a:1024099825458 · Cited by 541 · cited 1×
“Since the basic version of ROC is only suitable for binary classification, we use a multi-class modification of it [67]”
MethodAUC: a statistically consistent and more discriminating measure than accuracy
Ling, C. X., Huang, J., & Zhang, H. (2003) · Cited by 501
Training deep neural networks on imbalanced data sets
Wang, S., Liu, W., Wu, J., Cao, L., Meng, Q., & Kennedy, P. J. (2016) · 2016 International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn.2016.7727770 · Cited by 491 · cited 1×
“New kinds of loss function for neural networks training were also developed [27].”
BackgroundRelay Backpropagation for Effective Learning of Deep Convolutional Neural Networks
Shen, L., Lin, Z., & Huang, Q. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-46478-7_29 · Cited by 456 · cited 1×
“An oversampling approach specific to neural networks optimized with stochastic gradient descent is class-aware sampling [38].”
BackgroundChan, P. K., & Stolfo, S. J. (1998) · Cited by 452
Learning When Data Sets are Imbalanced and When Costs are Unequal and Unknown
Maloof, M. A. (2003) · Cited by 406
A novelty detection approach to classification
Japkowicz, N., Myers, C. E., & Gluck, M. A. (1995) · Cited by 392
Automated annotation of coral reef survey images
Beijbom, O., Edmunds, P. J., Kline, D. I., Mitchell, B. G., & Kriegman, D. J. (2012) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2012.6247798 · Cited by 347 · cited 1×
“There are plenty of examples in domains like computer vision [6, 7, 8, 9, 10], medical diagnosis [11, 12], fraud detection [13] and others [14, 15, 16] where this issue is highly significant and the frequency of one class (e.g., cancer) can be 1000 times less than another class (e.g., healthy…”
BackgroundCost-Sensitive Learning with Neural Networks.
Kukar, M., & Kononenko, I. (1998) · Cited by 244
Nonlinear Autoassociation Is Not Equivalent to PCA
Japkowicz, N., Hanson, S. J., & Gluck, M. A. (2000) · Neural Computation · doi:10.1162/089976600300015691 · Cited by 169 · cited 1×
“Then, the classification of a new example is made based on a reconstruction error between the input and output patterns, e.g. absolute error, squared sum of errors, Euclidean or Mahalanobis distance [46, 47, 48].”
MethodImageNet pre-trained models with batch normalizationPreprint
Simon, M., Rodner, E., & Denzler, J. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1612.01452 · Cited by 157 · cited 1×
“A model architecture employed for this dataset is ResNet-10 [60], i.e. a residual network [61] with batch normalization layers that are known to accelerate deep networks training [62].”
MethodWang, K., Makond, B., Chen, K.-H., & Wang, K.-M. (2013) · Applied Soft Computing · doi:10.1016/j.asoc.2013.09.014 · Cited by 140 · cited 1×
“It has been shown that oversampling is effective, yet it can lead to overfitting [33, 34].”
Background(2004) · Cognitive technologies · doi:10.1007/978-3-642-18859-6 · Cited by 128
Improving Minority Class Prediction Using Case-Specific Feature Weights
Cardie, C., & Nowe, N. (1997) · Smith ScholarWorks (Smith College) · Cited by 109
Neural Network Classification and Prior Class Probabilities
Lawrence, S., Burns, I., Back, A. D., Tsoi, A. C., & Giles, C. L. (2012) · Lecture notes in computer science · doi:10.1007/978-3-642-35289-8_19 · Cited by 100
Classification and knowledge discovery in protein databases
Predrag Radivojac, A., Chawla, N. V., Dunker, A. K., & Zoran Obradovic, A. (2004) · Journal of Biomedical Informatics · doi:10.1016/j.jbi.2004.07.008 · Cited by 91
Johnson, B. A., Tateishi, R., & Hoan, N. T. (2013) · International Journal of Remote Sensing · doi:10.1080/01431161.2013.810825 · Cited by 89
The problem of bias in training data in regression problems in medical decision support
Namee, B. M., Cunningham, P., Byrne, S., & Corrigan, O. I. (2002) · Artificial Intelligence in Medicine · doi:10.1016/s0933-3657(01)00092-6 · Cited by 88 · cited 1×
“There are plenty of examples in domains like computer vision [6, 7, 8, 9, 10], medical diagnosis [11, 12], fraud detection [13] and others [14, 15, 16] where this issue is highly significant and the frequency of one class (e.g., cancer) can be 1000 times less than another class (e.g., healthy…”
BackgroundOn the relation of performance to editing in nearest neighbor rules
Koplowitz, J., & Brown, T. A. (1981) · Pattern Recognition · doi:10.1016/0031-3203(81)90102-3 · Cited by 87 · cited 1×
“A more general approach than undersampling is data decontamination that can involve relabeling of some examples [41, 42].”
BackgroundCost-aware Pre-training for Multiclass Cost-sensitive Deep LearningPreprint
Chung, Y.-A., Lin, H.-T., & Yang, S. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1511.09337 · Cited by 84 · cited 2×
“Some previous studies showed results on cost sensitive learning of deep neural networks [24, 25, 26].”
ResultDetection of concealed cars in complex cargo X-ray imagery using Deep Learning
Jaccard, N., Rogers, T. W., Morton, E. J., & Griffin, L. D. (2017) · Journal of X-Ray Science and Technology · doi:10.3233/xst-16199 · Cited by 68 · cited 1×
“One of the most commonly used method in deep learning [16, 30, 31, 32].”
MethodAn Approach to Imbalanced Data Sets Based on Changing Rule Strength
Grzymala‐Busse, J. W., Goodwin, L., Grzymała-Busse, W. J., & Zheng, X. (2004) · Cognitive technologies · doi:10.1007/978-3-642-18859-6_21 · Cited by 65 · cited 1×
“There are plenty of examples in domains like computer vision [6, 7, 8, 9, 10], medical diagnosis [11, 12], fraud detection [13] and others [14, 15, 16] where this issue is highly significant and the frequency of one class (e.g., cancer) can be 1000 times less than another class (e.g., healthy…”
BackgroundThe Novelty Detection Approach for Different Degrees of Class Imbalance
Lee, H., & Cho, S. (2006) · Lecture notes in computer science · doi:10.1007/11893257_3 · Cited by 51 · cited 1×
“This method has proved to work well for extremely high imbalance when classification problem turns info anomaly detection [49].”
MethodTowards Effective Classification of Imbalanced Data with Convolutional Neural Networks
Raj, V., Magg, S., & Wermter, S. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-46182-3_13 · Cited by 44 · cited 1×
“Some previous studies showed results on cost sensitive learning of deep neural networks [24, 25, 26].”
ResultKing, I., Wang, J., Wang, D., & Chan, L. (2006) · Lecture notes in computer science · doi:10.1007/11893295 · Cited by 30
Restricted Decontamination for the Imbalanced Training Sample Problem
Barandela, R., Rangel, E., Sánchez, J. S., & Ferri, F. J. (2003) · Lecture notes in computer science · doi:10.1007/978-3-540-24586-5_52 · Cited by 27
The iNaturalist Species Classification and Detection DatasetPreprint
Horn, G. V., Aodha, O. M., Song, Y., Cui, Y., Sun, C., Shepard, A., Adam, H., Perona, P., & Belongie, S. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1707.06642 · Cited by 18
Novelty Detection Using Auto-Associative Neural Network
Sohn, H., Worden, K., & Farrar, C. R. (2001) · Dynamic Systems and Control · doi:10.1115/imece2001/dsc-24571 · Cited by 16 · cited 1×
“absolute error, squared sum of errors, Euclidean or Mahalanobis distance [46, 47, 48].”
Background
Cited by
Sorted by citations
A survey on Image Data Augmentation for Deep Learning
Shorten, C., & Khoshgoftaar, T. M. (2019) · Journal of Big Data · doi:10.1186/s40537-019-0197-0 · Cited by 13,203
Class-Balanced Loss Based on Effective Number of Samples
Cui, Y., Jia, M., Lin, T.-Y., Song, Y., & Belongie, S. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00949 · Cited by 3,094
“CNNs trained on such data perform poorly for weakly represented classes [19, 15, 41, 4].”
BackgroundSurvey on deep learning with class imbalance
Johnson, J. M., & Khoshgoftaar, T. M. (2019) · Journal of Big Data · doi:10.1186/s40537-019-0192-5 · Cited by 3,032
CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep LearningPreprint
Rajpurkar, P., Irvin, J., Zhu, K., Yang, B., Mehta, H., Duan, T., Ding, D. Y., Bagul, A., Langlotz, C. P., Shpanskaya, K. S., et al. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1711.05225 · Cited by 4,226
“We train the model using minibatches of size 16, where we oversample the minority (positive) class (Buda et al., 2017).”
MethodLearning Imbalanced Datasets with Label-Distribution-Aware Margin LossPreprint
Cao, K., Wei, C., Gaidon, A., Aréchiga, N., & Ma, T. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1906.07413 · Cited by 1,140
“, [46, 60, 5, 6] and references therein) and under-sampling the frequent classes (see, e.”
BackgroundMethodInfluential
Metadata
- Authors
- Buda, M., Maki, A., & Mazurowski, M. A.
- Affiliations
- Duke University (US)
- KTH Royal Institute of Technology (SE)
- Corresponding author
- Buda, Mateusz
- Venue
- Neural Networks
- Publisher
- Elsevier BV
- Published
- 2018-10
- Volume
- 106
- Pages
- 249-259
- ISSN
- 0893-6080 (ISSN-L)
- OpenAlex
- W2767106145
- Cited by
- 3,088
- References
- 83
- Citations by year
- 2026: 459 · 2025: 507 · 2024: 371 · 2023: 405 · 2022: 405 · 2021: 380 · 2020: 308 · 2019: 189 · 2018: 58
- FWCI
- 131.58
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- Neural Networks (publishedVersion)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- KTH Publication Database DiVA (KTH Royal Institute of Technology) (submittedVersion, other-oa)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Imbalanced Data Classification Techniques
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning and Data Classification
- Keywords
- class imbalance, convolutional neural networks, oversampling, undersampling, two-stage training, thresholding, area under the receiver operating characteristic, multi-class classification, overfitting, MNIST, CIFAR-10, ImageNet
- MeSH terms
- Machine Learning, Humans, Probability, ROC Curve, Neural Networks, Computer
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
