Graph Regularization Methods in Soft Detector Fusion

Abstract

This paper presents a theoretical derivation of two new graph-based regularization methods for fusing the individual results of multiple detectors (two-class classifiers). The proposed approach considers linear combination of the individual detector statistics and its extension to a general nonlinear fusion method known as α-integration. A cost function that includes a mean-square error and a regularization term is minimized. The inclusion of the regularization term, which is based on graph signal processing, reduces the dispersion of the fused statistics, and thus improves the separation between the fused statistics corresponding to every detection hypothesis. The proposed methods (linear and non-linear regularized α-integration) are experimentally compared with commonly used classification methods (random forest, linear and quadratic discriminant analysis, and naive Bayes) and competitive fusion methods (Dempster-Shafer, copulas, behavior knowledge space, independent component analysis mixture modeling, majority voting, the mean, and αintegration). Two challenging problems were approached using simulated and electroencephalographic data, respectively: (i) detection of ultrasound pulses buried in high noise, and (ii) detection of changes in electroencephalographic signals for neuropsychological test staging. An experimental convergence analysis of the proposed regularized method for these two applications is included. Besides, the proposed methods were tested using several benchmark datasets. Results on the basis of classification accuracy, kappa index, F1 score, and receiver operating characteristic curve analysis demonstrate the superiority of the proposed regularized fusion methods.

Image Deraining Algorithm Based on Multi-Scale Features

“…Early fusion is the best option with model knowledge given that both early and late fusion degrade due to a finite training set. This degradation is shown to be greater for early fusion due to the dimensionality increase in the feature space, so, eventually, late fusion could be a better option in a practical setting [35][36][37][38].…”

A Comparative Study on Recent Automatic Data Fusion Methods

“…) is ultimately a function of multivariate random variables x 1 , x 2 , which will generally be different from 1) and therefore rule (2) can never achieve a probability of error lower than rule (1), assuming knowledge of P(k = 1/x 1 , x 2 ). For instance, let us analyze the case of the fusion method based on α-integration [25][26][27][28]. First, we apply Bayes' rule:…”

Hyperspectral Image Classification Based on Adaptive Global–Local Feature Fusion

“…HSI classification distinguishes the corresponding categories of each pixel, which is a basic and key application technology in remote sensing and which can be successfully utilized in numerous fields such as mineral detection, environment detection, and crop monitoring [4][5][6]. Early applied HSI classification methods, including random forest [7], support vector machine (SVM) [8], and graph-based [9] methods, enhanced feature classification ability by exploring rich and effective spectral information. Other models, such as independent component analysis [10] and principal component analysis (PCA) [11], are often used to identify valid spectral features.…”