Model observer performance, computed theoretically using cascaded systems analysis (CSA), was compared to the performance of human observers in detection and discrimination tasks. Dual-energy (DE) imaging provided a wide range of acquisition and decomposition parameters for which observer performance could be predicted and measured. This work combined previously derived observer models (e.g., Fisher-Hotelling and non-prewhitening) with CSA modeling of the DE image noise-equivalent quanta (NEQ) and imaging task (e.g., sphere detection, shape discrimination, and texture discrimination) to yield theoretical predictions of detectability index and area under the receiver operating characteristic . Theoretical predictions were compared to human observer performance assessed using 9-alternative forced-choice tests to yield measurement of as a function of DE image acquisition parameters (viz., allocation of dose between the low- and high-energy images) and decomposition technique [viz., three DE image decomposition algorithms: standard log subtraction (SLS), simple-smoothing of the high-energy image (SSH), and anti-correlated noise reduction (ACNR)]. Results showed good agreement between theory and measurements over a broad range of imaging conditions. The incorporation of an eye filter and internal noise in the observer models demonstrated improved correspondence with human observer performance. Optimal acquisition and decomposition parameters were shown to depend on the imaging task; for example, ACNR and SSH yielded the greatest performance in the detection of soft-tissue and bony lesions, respectively. This study provides encouraging evidence that Fourier-based modeling of NEQ computed via CSA and imaging task provides a good approximation to human observer performance for simple imaging tasks, helping to bridge the gap between Fourier metrics of detector performance (e.g., NEQ) and human observer performance.
The authors thank Drs. R. Van Metter and J. Yorkston (Carestream Health Inc., Rochester, NY) for stimulating discussions and collaboration in creating the DE imaging prototype. Thanks also to D. J. Tward, C. A. Varon, and H. Kashani (Ontario Cancer Institute, Toronto, ON) for assistance with the experiments and analysis. This work was supported by a Cunningham Fellowship Award, a University of Toronto Open Scholarship Award, and National Institute of Health Grant No. R01-CA-112163.
II. THEORETICAL METHODS
II.A. DE imaging performance modeling
II.A.1. Generalized DE image decomposition
II.A.2. Cascaded systems analysis of DE imaging
II.B. Model observers and the detectability index
II.B.1. The Fisher-Hotelling observer (FH)
II.B.2. The Fisher-Hotelling observer with eye filter (FHE)
II.B.3. The Non-prewhitening observer (NPW)
II.B.4. The non-prewhitening observer with eye filter (NPWE)
II.C. Imaging tasks
II.C.1. Sphere detection task
II.C.2. Shape discrimination task
II.C.3. Texture discrimination task
II.D. Comparison of with human observer performance
III. EXPERIMENTAL METHODS
III.A. Imaging setup
III.A.1. DE imaging system
III.A.2. Task phantom
III.B. Human observer study
III.B.1. Multiple-alternative forced choice (MAFC) tests
III.B.2. Dependence of performance on acquisition technique
III.B.3. Dependence of performance on decomposition technique
III.B.4. Experimental setup
IV.A. Dependence of performance on acquisition technique
IV.B. Dependence of performance on decomposition technique
V. DISCUSSION AND CONCLUSION
Data & Media loading...
Article metrics loading...
Full text loading...