Receiver Operating Characteristic Curve Confidence Intervals and Regions

Kerekes, John P.

doi:10.1109/lgrs.2008.915928

Cited by 179 publications

(65 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, one method for the calculation of the standard error is where b and c represent the frequency of the discordant pairs (the cases for which the two classifiers compared differed in labelling, which lie in the off-diagonal elements of the 2x2 confusion matrix used in the McNemar test) in the sample of size n (Newcombe, 1998); although there are concerns with methods of estimating the confidence interval for use in association with a McNemar test (Lloyd, 1990;Newcombe, 1998). Confidence intervals may also be fitted for use in relation to other tests used in comparing classifications derived from remotely sensed data such as the receiver operating characteristics (ROC) curve (Kerekes, 2008).…”

Section: Confidence Intervalsmentioning

confidence: 99%

Classification accuracy comparison: Hypothesis tests and the use of confidence intervals in evaluations of difference, equivalence and non-inferiority

Foody

2009

Remote Sensing of Environment

199

138

View full text Add to dashboard Cite

show abstract

Section: Confidence Intervalsmentioning

confidence: 99%

Classification accuracy comparison: Hypothesis tests and the use of confidence intervals in evaluations of difference, equivalence and non-inferiority

Foody

2009

Remote Sensing of Environment

199

138

View full text Add to dashboard Cite

show abstract

“…For the RSLAD method, the number of sampled pixels p, the projected dimension K and the residual threshold ε on the PaviaC dataset are manually set to be 150, 50 and 10 −9 , respectively; the p, K and ε on the San Diego dataset are manually set to be 120, 50 and 1.7 × 10 −10 , respectively; the p, K and ε on the Botswana dataset are manually set to be 100, 50 and 4 × 10 −10 , respectively; and the p, K and ε on the HyMap dataset are manually set to be 200, 60 and 1.7 × 10 −10 , respectively. Figure 7 illustrates the ROC curves and confidence intervals and regions [51] of RSLAD and other four methods on the four datasets. For the PaviaC dataset of Figure 7a, RSLAD has the lowest false alarm rate at 100% probability of detection.…”

Section: Detection Performance the Rslad Methodsmentioning

confidence: 99%

A Randomized Subspace Learning Based Anomaly Detector for Hyperspectral Imagery

Sun

Tian

et al. 2018

Remote Sensing

View full text Add to dashboard Cite

This paper proposes a randomized subspace learning based anomaly detector (RSLAD) for hyperspectral imagery (HSI). Improved from robust principal component analysis, the RSLAD assumes that the background matrix is low-rank, and the anomaly matrix is sparse with a small portion of nonzero columns (i.e., column-wise). It also assumes the anomalies do not lie in the column subspace of the background and aims to find a randomized subspace of the background to detect the anomalies. First, random techniques including random sampling and random Hadamard projections are implemented to construct a coarse randomized columns subspace of the background with reduced computational cost. Second, anomaly columns are searched and removed from the coarse randomized column subspace by solving a series of least squares problems, resulting in a purified randomized column subspace. Third, the nonzero columns in the anomaly matrix are located by projecting all the pixels on the orthogonal subspace of the purified subspace, and the anomalies are finally detected based on the L 2 norm of the columns in the anomaly matrix. The detection performance of RSLAD is compared with four state-of-the-art methods, including global Reed-Xiaoli (GRX), local RX (LRX), collaborative-representation based detector (CRD), and low-rank and sparse matrix decomposition base anomaly detector (LRaSMD). Experimental results show good detection performance of RSLAD with lower computational cost. Therefore, the proposed RSLAD offers an alternative option for hyperspectral anomaly detection.

show abstract

“…In this subsection, the detection performance is evaluated quantitatively by receiver operating characteristic (ROC) curves [41] and the normalized background-anomaly separation maps. Based on the ground truth, the ROC curve plots the relationship between the detection rate DR and the false alarm rate FAR ; DR and FAR are defined as follows: is the total number of pixels in the image.…”

Section: Detection Performancementioning

confidence: 99%

Hyperspectral Anomaly Detection via Background Estimation and Adaptive Weighted Sparse Representation

Zhu

Wen

2018

Remote Sensing

View full text Add to dashboard Cite

Abstract:Anomaly detection is an important task in hyperspectral imagery (HSI) processing. It provides a new way to find targets that have significant spectral differences from the majority of the dataset. Recently, the representation-based methods have been proposed for detecting anomaly targets in HSIs. It is essential for this type of method to construct a valid background dictionary to distinguish anomaly and background accurately. In this paper, a novel hyperspectral anomaly detection method based on background estimation and adaptive weighted sparse representation has been proposed. Firstly, to obtain the effective background dictionary without anomaly information, a new background dictionary construction strategy is designed. Secondly, the sparse representation based on the constructed background dictionary is utilized on the dataset. Anomalies and background are distinguished through the response of the residual matrix. Thirdly, the residual matrix is weighted adaptively from global and local domains, which makes anomalies and background more discriminative. An important advantage of the proposed method is that it considers the properties of anomalies in both spectral and spatial domains. Experiments on three HSI datasets reveal that our proposed method achieves an outstanding detection performance compared with the other anomaly detection algorithms.

show abstract

Receiver Operating Characteristic Curve Confidence Intervals and Regions

Cited by 179 publications

References 13 publications

Classification accuracy comparison: Hypothesis tests and the use of confidence intervals in evaluations of difference, equivalence and non-inferiority

Classification accuracy comparison: Hypothesis tests and the use of confidence intervals in evaluations of difference, equivalence and non-inferiority

A Randomized Subspace Learning Based Anomaly Detector for Hyperspectral Imagery

Hyperspectral Anomaly Detection via Background Estimation and Adaptive Weighted Sparse Representation

Contact Info

Product

Resources

About