Pre-Conference Talk by LIU Zhonghang | Rethinking Unsupervised Outlier Detection via Multiple Thresholding

In the realm of current 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 hinders some real applications of current outlier detectors but also causes these methods not able 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 unlabeled 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. 

This is a Pre-Conference talk for The European Conference on Computer Vision (ECCV 2024).
 

Zhonghang LIU is a PhD candidate under the supervision of Prof. Daniel LIN. His research interest is machine learning and computer vision, specifically in unsupervised outlier detection for images.