Cost-Sensitive Learning of Deep Feature Representations From Imbalanced Data

• Published in: IEEE transaction on neural networks and learning systems Vol. 29; no. 8; pp. 3573 - 3587
• Main Authors: Khan, Salman H., Hayat, Munawar, Bennamoun, Mohammed, Sohel, Ferdous A., Togneri, Roberto
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.08.2018; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Artificial neural networks; Australia; Classification; Classifiers; Computer applications; Computer vision; Convolutional neural networks (CNNs); cost-sensitive (CoSen) learning; data imbalance; Data sampling; Image classification; loss functions; Neural networks; Object recognition; Representations; Tag clouds; Testing; Training; Training data
• DOI: 10.1109/TNNLS.2017.2732482
• ISSN: 2162-237X; 2162-2388

Class imbalance is a common problem in the case of real-world object detection and classification tasks. Data of some classes are abundant, making them an overrepresented majority, and data of other classes are scarce, making them an underrepresented minority. This imbalance makes it challenging for a classifier to appropriately learn the discriminating boundaries of the majority and minority classes. In this paper, we propose a cost-sensitive (CoSen) deep neural network, which can automatically learn robust feature representations for both the majority and minority classes. During training, our learning procedure jointly optimizes the class-dependent costs and the neural network parameters. The proposed approach is applicable to both binary and multiclass problems without any modification. Moreover, as opposed to data-level approaches, we do not alter the original data distribution, which results in a lower computational cost during the training process. We report the results of our experiments on six major image classification data sets and show that the proposed approach significantly outperforms the baseline algorithms. Comparisons with popular data sampling techniques and CoSen classifiers demonstrate the superior performance of our proposed method.