# One-Class Feature Learning Using Intra-Class Splitting

## TL;DR

This paper introduces a new one-class feature learning approach that splits normal class data into typical and atypical samples, enabling effective feature extraction without multi-class pretraining, and demonstrates superior performance on image datasets.

## Contribution

The proposed intra-class splitting method allows for effective feature learning in one-class classification without relying on multi-class pretraining, a novel approach in the field.

## Key findings

- Outperformed baseline models on three image datasets
- Effective in extracting valuable features for one-class classification
- Demonstrated robustness across different datasets

## Abstract

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.