Generalization Evaluation of Machine Learning Numerical Observers for Image Quality Assessment

Author

Kalayeh, M.M. ; Marin, T. ; Brankov, J.G.

Author_Institution

Dept. of Electr. & Comput. Eng., Illinois Inst. of Technol., Chicago, IL, USA

Volume

60

Issue

3

fYear

2013

fDate

Jun-13

Firstpage

1609

Lastpage

1618

Abstract

In this paper, we present two new numerical observers (NO) based on machine learning for image quality assessment. The proposed NOs aim to predict human observer performance in a cardiac perfusion-defect detection task for single-photon emission computed tomography (SPECT) images. Human observer (HumO) studies are now considered to be the gold standard for task-based evaluation of medical images. However such studies are impractical for use in early stages of development for imaging devices and algorithms, because they require extensive involvement of trained human observers who must evaluate a large number of images.

Keywords

learning (artificial intelligence); single photon emission computed tomography; SPECT images; cardiac perfusion-defect detection task; gold standard for task-based evaluation; human observer performance; human observer study; image quality assessment; imaging algorithms; imaging devices; machine learning numerical observers; medical images; single-photon emission computed tomography images; Computational modeling; Feature extraction; Kernel; Noise; Numerical models; Observers; Support vector machines; Channelized hotelling observer (CHO); human observer (HumO); image quality; numerical observer; relevance vector machine (RVM); single-photon emission computed tomography (SPECT);

fLanguage

English

Journal_Title

Nuclear Science, IEEE Transactions on

Publisher

ieee

ISSN

0018-9499

Type

jour

DOI

10.1109/TNS.2013.2257183

Filename

6515707