One-Class Feature Learning Using Intra-Class Splitting
This addresses the problem of feature learning in one-class classification for scenarios with limited data, offering a novel approach that eliminates the need for reference multi-class datasets, though it appears incremental as it builds on existing splitting concepts.
The paper tackles the challenge of feature learning in one-class classification, where only samples from a single class are available, by proposing a method that splits the normal class into typical and atypical subsets and uses closeness and dispersion losses for joint training, resulting in outperformance of baseline models on three image datasets.
This paper proposes a novel generic one-class feature learning method based on intra-class splitting. In one-class classification, feature learning is challenging, because only samples of one class are available during training. Hence, state-of-the-art methods require reference multi-class datasets to pretrain feature extractors. In contrast, the proposed method realizes feature learning by splitting the given normal class into typical and atypical normal samples. By introducing closeness loss and dispersion loss, an intra-class joint training procedure between the two subsets after splitting enables the extraction of valuable features for one-class classification. Various experiments on three well-known image classification datasets demonstrate the effectiveness of our method which outperformed other baseline models in average.