Robust Principal Component Analysis in Imaging: Novel Algorithms for Outlier Detection and Precision

Robust Principal Component Analysis (RPCA) has emerged as one of the most influential techniques for robust dimensionality reduction, feature extraction, denoising, and low-rank representation in imaging and signal processing. As a robust extension of classical Principal Component Analysis (PCA), RPCA addresses the limitations of conventional PCA, which is highly sensitive to outliers, non-Gaussian noise, and corrupted observations. While numerous dimensionality reduction methods—including Independent Component Analysis (ICA), Linear Discriminant Analysis (LDA), Non-negative Matrix Factorization (NMF), Kernel PCA (KPCA), manifold learning approaches, and deep learning-based feature extraction—have been developed for high-dimensional data analysis, RPCA remains particularly attractive because of its ability to recover the intrinsic low-rank structure and isolate sparse anomalies simultaneously. These capabilities have led to widespread applications in hyperspectral imaging, medical imaging, remote sensing, industrial inspection, computer vision, image restoration, background subtraction, and anomaly detection.Despite significant progress, existing RPCA approaches continue to face important challenges. Classical methods based on Principal Component Pursuit (PCP) or sparse L1-norm optimization often incur high computational costs. They may struggle with large-scale datasets, structured noise, dynamic backgrounds, or complex image variability. More recent approaches based on non-convex Lp-norm optimization (0 < p < 1) and L0-norm-inspired sparse representations have shown improved robustness and reconstruction accuracy. Nevertheless, challenges remain in balancing computational efficiency, optimization stability, scalability, interpretability, and reliable outlier detection across diverse and complex imaging applications.This Research Topic invites original contributions that advance the theory, algorithms, and applications of Robust Principal Component Analysis in imaging. We welcome research on novel RPCA models, optimization techniques, and hybrid frameworks that improve robustness to outliers, noise, missing data, and complex image degradations while enhancing reconstruction accuracy, computational efficiency, and interpretability. Contributions that integrate RPCA with other dimensionality reduction methods, machine learning, deep learning, or multimodal imaging are also encouraged.Topics of interest include, but are not limited to:-- Robust Principal Component Analysis (RPCA) and Principal Component Pursuit (PCP)-- Sparse and low-rank matrix and tensor decomposition-- Robust dimensionality reduction and feature extraction-- Outlier detection, anomaly detection, and background modeling-- Sparse, graph-based, kernel, and nonlinear RPCA-- Deep learning-enhanced and hybrid RPCA methods-- Optimization algorithms for robust low-rank recovery-- Image denoising, restoration, reconstruction, and compression-- RPCA for hyperspectral, multispectral, medical, industrial, and remote sensing imaging-- Computationally efficient and scalable RPCA for high-dimensional data-- Explainable and interpretable robust learning methods-- Benchmark datasets, comparative evaluations, and real-world imaging applicationsSubmissions presenting theoretical developments, optimization strategies, benchmark studies, and practical imaging applications are particularly encouraged. Contributions that improve the robustness, precision, scalability, and interpretability of RPCA while addressing the challenges of modern imaging systems will help advance the next generation of intelligent image analysis and signal processing technologies. Topic editors: Stanley Abhadiomhen, Ernest Domanaanmwi Ganaa, Madhusudan Lanjewar Source: https://www.frontiersin.org/research-topics/73401/robust-principal-component-analysis-in-imaging-novel-algorithms-for-outlier-detection-and-precision