CGAN-plankton: Towards Large-Scale Imbalanced Class Generation and Fine-Grained Classification.

Chao Wang,Zhibin Yu,Haiyong Zheng,Nan Wang,Bing Zheng
DOI: https://doi.org/10.1109/icip.2017.8296402
2017-01-01
Abstract:Plankton classification is becoming critically important as people concentrate more on oceans and global environment changing. Data of plankton species naturally exhibit imbalance in their class distribution. Meanwhile, it arouses fine-grained classification challenge. Although Convolutional Neural Networks (CNNs) have human-level performance on image classification task, they tend to be biased to large classes without considering the imbalance issue. In this paper, we introduce Generative Adversarial Network (GAN) based generative model to overcome these challenges. Our proposed model consists of fully convolutional layers, and includes three parts: a generative model G, a discriminative model D and a classification model C. We train generative and discriminative models on small classes data to learn a discriminative features through D model and reduce mode missing problem to some extent. We implement classification task using shared CNN layers of D model on whole data. Experimental results show that our model significantly improved the F1 score on an imbalanced plankton dataset with well-generated plankton images.
What problem does this paper attempt to address?