Unsupervised Outlier Detection using Random Subspace and Subsampling Ensembles of Dirichlet Process Mixtures

Probabilistic mixture models are acknowledged as a valuable tool for unsupervised outlier detection owing to their interpretability and intuitive grounding in statistical principles. Within this framework, Dirichlet process mixture models emerge as a compelling alternative to conventional finite mixture models for both clustering and outlier detection tasks. However, despite their evident advantages, the widespread adoption of Dirichlet process mixture models in unsupervised outlier detection has been hampered by challenges related to computational inefficiency and sensitivity to outliers during the construction of detectors. To tackle these challenges, we propose a novel outlier detection method based on ensembles of Dirichlet process Gaussian mixtures. The proposed method is a fully unsupervised algorithm that capitalizes on random subspace and subsampling ensembles, not only ensuring efficient computation but also enhancing the robustness of the resulting outlier detector. Moreover, the proposed method leverages variational inference for Dirichlet process mixtures to ensure efficient and fast computation. Empirical studies with benchmark datasets demonstrate that our method outperforms existing approaches for unsupervised outlier detection.

Single-shot refractive index slice imaging using spectrally multiplexed optical transfer function reshaping

The refractive index (RI) of cells and tissues is crucial in pathophysiology as a noninvasive and quantitative imaging contrast. Although its measurements have been demonstrated using three-dimensional quantitative phase imaging methods, these methods often require bulky interferometric setups or multiple measurements, which limits the measurement sensitivity and speed. Here, we present a single-shot RI imaging method that visualizes the RI of the in-focus region of a sample. By exploiting spectral multiplexing and optical transfer function engineering, three color-coded intensity images of a sample with three optimized illuminations were simultaneously obtained in a single-shot measurement. The measured intensity images were then deconvoluted to obtain the RI image of the in-focus slice of the sample. As a proof of concept, a setup was built using Fresnel lenses and a liquid-crystal display. For validation purposes, we measured microspheres of known RI and cross-validated the results with simulated results. Various static and highly dynamic biological cells were imaged to demonstrate that the proposed method can conduct single-shot RI slice imaging of biological samples with subcellular resolution.