CVLGDec 8, 2020

Nonlinear Cook distance for Anomalous Change Detection

arXiv:2012.12307v1
AI Analysis

This work addresses the problem of anomalous change detection in remote sensing images, offering an incremental improvement to existing chronochrome approaches.

This paper proposes a method to detect anomalous changes in remote sensing images by identifying influential points using a nonlinear extension of Cook distance. The method leverages random Fourier features for efficient nonlinear impact measurement, demonstrating good empirical performance visually and quantitatively with ROC curves on multispectral images.

In this work we propose a method to find anomalous changes in remote sensing images based on the chronochrome approach. A regressor between images is used to discover the most {\em influential points} in the observed data. Typically, the pixels with largest residuals are decided to be anomalous changes. In order to find the anomalous pixels we consider the Cook distance and propose its nonlinear extension using random Fourier features as an efficient nonlinear measure of impact. Good empirical performance is shown over different multispectral images both visually and quantitatively evaluated with ROC curves.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes