Model for the detection of signals in images with multiple suspicious locations

Abstract

A signal detection model is presented that combines a signal model and a noise model providing mathematical descriptions of the frequency of appearance of the signals, and of the signal-like features naturally occurring in the background. We derive expressions for the likelihood functions for the whole ensemble of observed suspicious locations, in various possible combinations of signals and false signal candidates. As a result, this formalism is able to describe several new types of detection tests using likelihood ratio statistics. We have a global image abnormality test and an individual signal detection test. The model also provides an alternative mechanism in which is selected the combination of signal and noise features candidates that has the maximum likelihood. These tests can be analyzed with a variety of operating characteristic curves ͑ROC, LROC, FROC, etc.͒. In the mathematical formalism of the model, all the details characterizing the suspicious features are reduced to a single scalar function, which we name the signal specificity function, representing the frequency that a signal takes a certain value relative to the frequency of having a false signal with the same value in an image of given size. The signal specificity function ranks the degree of suspiciousness of the features found, and can be used to unify into a single score all the suspicious feature characteristics, and then apply the usual decision conventions as in the Swensson's detection model ͓Med. Phys. 23, 1709-1725 ͑1996͔͒. We present several examples in which these tests are compared. We also show how the signal specificity function can be used to model various degrees of accuracy of the observer's knowledge about image noise and signal statistical properties. Aspects concerning modeling of the human observer are also discussed.

Objective assessment of image quality and dose reduction in CT iterative reconstruction

“…Due to the limitations of SKE tasks, recent studies examine the possibility of allowing for unknown signal location. 22,[25][26][27][28][29][30] For forcedchoice experiments, the figure of merit remains the percentage correct. More generally, the probability of correct joint detection and localization can be plotted vs the probability of false positives; the figure of merit is the area under this curve.…”

Nonparametric signal detectability evaluation using an exponential transformation of the FROC curve

“…In order to test the EFROC estimator and to compare it with its nonparametric counterparts for ROC, LROC, and AFROC analyses, we carried out a simulation study using a previously presented model, 34 with parameterization derived from work involving searching for signals on PET images. 6 The model considers small signals of size comparable to the spatial resolution of the imaging device so that their contrast relative to the surrounding background is practically the only measurable variable indicative to their degree of suspiciousness.…”

Evaluation of diagnostic accuracy in free-response detection-localization tasks using ROC tools

“…Proper quantification of the diagnostic accuracy in the FROC data is not trivial and requires specialized tools (Chapter 9.3 in [1], Chapter 8 in [11], [12]). The fundamental tool in the FROC analysis is the FROC curve 5,[13][14][15] which reflects the tradeoff between the false-positives and true-positive findings with changing decision threshold. Some FROC approaches rely directly on the FROC curve and associated indices, [16][17][18] while others focus on specific summary indices which do not necessarily agree with the empirical FROC curve [11(Ch.…”