Rethinking Unsupervised Outlier Detection via Multiple Thresholding
This work addresses a bottleneck in unsupervised outlier detection for image data, offering an incremental improvement to enhance existing methods.
The paper tackles the problem of determining optimal thresholds for outlier scores in unsupervised image outlier detection, which is ill-posed and limits real-world applications and self-supervised enhancement. The proposed multiple thresholding (Multi-T) module significantly improves existing scoring methods and enables a naive distance-based method to achieve state-of-the-art performance.
In the realm of unsupervised image outlier detection, assigning outlier scores holds greater significance than its subsequent task: thresholding for predicting labels. This is because determining the optimal threshold on non-separable outlier score functions is an ill-posed problem. However, the lack of predicted labels not only hiders some real applications of current outlier detectors but also causes these methods not to be enhanced by leveraging the dataset's self-supervision. To advance existing scoring methods, we propose a multiple thresholding (Multi-T) module. It generates two thresholds that isolate inliers and outliers from the unlabelled target dataset, whereas outliers are employed to obtain better feature representation while inliers provide an uncontaminated manifold. Extensive experiments verify that Multi-T can significantly improve proposed outlier scoring methods. Moreover, Multi-T contributes to a naive distance-based method being state-of-the-art.