Unified SNR analysis of medical imaging systems

Wagner, Robert F.; Brown, David G.

doi:10.1088/0031-9155/30/6/001

Cited by 343 publications

(221 citation statements)

References 59 publications

Supporting

Mentioning

218

Contrasting

Unclassified

Order By: Relevance

“…We assume that d' is proportional to the image contrast-to-noise ratio, Clan' where C is the contrast of the object as compared with background and an is the standard deviation ofthe display noise. The Rose (1973) model for a disk and the ideal observer model for a given object both have this form (Wagner & Brown, 1985). In an X-ray system that maintains a constant gray level, a; is inversely related to the number of X-ray quanta and hence X-ray dose, D , and we have…”

Section: Preliminary Experimental Resultsmentioning

confidence: 99%

See 1 more Smart Citation

An adaptive reference/test paradigm: Application to pulsed fluoroscopy perception

Xue

Thomas

Gilmore

et al. 1998

Behavior Research Methods, Instruments, & Computers

View full text Add to dashboard Cite

Wedeveloped an adaptive forced-choice method whereby reference and test presentations were alternated in order to minimize effects from variables such as subject attention level. In our demonstration example of an X-rayfluoroscopy perception study, we measured detectability of low-contrast objects in noisy image sequences and determined X-raydose levels for equivalent detectability of identical contrasts for a new test acquisition method (fluoroscopy at 15 acq/sec) as compared with a reference (conventional fluoroscopy at 30acq/sec). In preliminary experiments, we found a dose savings with the test method. We derived parameter uncertainties for the adaptive procedure and demonstrated their applicability with Monte Carlo simulations. Repeated experiments on a single subject demonstrated reduced standard errors due to the reduction of day-to-day variations. It is believed that the method can be applied in a variety of situations in which one needs to compare perception measurements.We have developed a new experimental method that addresses issues in human perception studies. First, a subject's response can vary due to fatigue, a lapse in attention, or possible physiological changes. Kahneman (1973) discussed the importance of attention in perception experiments, indicating that attention may affect the sensory mechanism as well as the decision mechanism. Taylor (1967) observed variation in performance from one day to the next, and others (Hall, 1981;Madigan & Williams, 1987) reported variations within single experiments. This variation complicates comparison of results. For example, in our experiments in medical X-ray fluoroscopy, we compared new test acquisition methods with an established reference technique. Some of the effects that we wished to measure are comparable to day-to-day variations, and our solution was to alternate reference and test presentations over a time scale of seconds and obtain responses under identical conditions. Second, variability between subjects also presents a dilemma. For an independent variable such as image contrast, many trial-and-error presentations are required to arrive at display conditions that match the floor and ceiling responses of all subjects. This problem is well known, and adaptive techniques can solve the problem (

show abstract

Section: Preliminary Experimental Resultsmentioning

confidence: 99%

“…For example, we use the reference/test method to evaluate new acquisition and image processing (Wilson, Jabri, Xue, & Aufrichtig, 1996) methods. There are numerous similar applications in medical imaging (Wagner & Brown, 1985) and image compression (Hearty, 1993). Many other applications exist in perception.…”

Section: Day Numbermentioning

confidence: 99%

An adaptive reference/test paradigm: Application to pulsed fluoroscopy perception

Xue

Thomas

Gilmore

et al. 1998

Behavior Research Methods, Instruments, & Computers

View full text Add to dashboard Cite

show abstract

“…The Macintosh TIFF images were converted to portable bit-mapped images using XV version 3.00 on the DECstation. The DECstation image analysis programs were written in C. Lesion detectability: The matched jilter method Considerable work has been done on detectability of low-contrast cylindrical targets in ultrasound (Hall et al 1993;Insana and Hall 1994;Lopez et al 1990Lopez et al , 1992Smith et al 1983;Thijssen et al 1988;Wagner and Brown 1985;Wagner et al 1983). The concept of the "matched filter" has been developed for ultrasound Wagner et al 1983) and has been used with phantoms containing low-contrast cylindrical targets to do objective contrast-detail studies (Hall et al 1993;Lopez et al 1992).…”

Section: Automation Methodsologymentioning

confidence: 99%

“…A pattern recognition algorithm, called the "matched filter method" (Wagner et al , 1985, has been modified for use in our detectability algorithm. Thus, a lesion signal-to-noise ratio ( SZVR)e, which depends on the expected size of the lesion searched for, is computed for each pair of pixel coordinates in digitized images of the phantoms.…”

Section: Introductionmentioning

confidence: 99%

Phantoms and automated system for testing the resolution of ultrasound scanners

Rownd

Madsen

Zagzebski

et al. 1997

Ultrasound in Medicine & Biology

View full text Add to dashboard Cite

Abstract-Tissue-mimicking phantoms and an automated system have been developed for testing the resolution performance of ultrasound scanners by determining detectability of low to higher contrast spherical lesions over the entire depth of field. Axial, lateral and elevational resolutions are accounted for simuRaneously and equally. Tissue-mimicking spherical simulated lesions are either 3 or 4 mm in diameter and have one of four different intrinsic material contrasts. For each diameter and contrast, there is a set of 109 lesions in a regular array with coplanar centers extending from 0.5-15.5 cm in depth. With the scan slice superimposed on the spheres, the image is frame-grabbed for automated analysis. A diameter-dependent lesion signal-to-noise ratio is computed for each pixel position in the image, exchrdhtg a 5-mm boundary. Two universal thresholds, resulting from maximization of agreement between the automated system and human observers, give rise to a depth range, or "resolution zone", over which detection exists for each type lesion. 0 1997 World Federation for Ultrasound in Medicine & Biology.

show abstract

“…It is most often integrated with other metrics to assess image quality for specific tasks, 6 , 7 , 8 , 9 and has been commonly utilized in the development, characterization, optimization, and comparison of many new imaging technologies such as computed radiography, (10) digital mammography, 11 , 12 storage phosphors for dental X‐ray, (13) and other devices in a preclinical 14 , 15 and clinical 16 , 17 environment.…”

Section: Introductionmentioning

confidence: 99%

Practical considerations for noise power spectra estimation for clinical CT scanners

Dolly

Chen

Anastasio

et al. 2016

J Applied Clin Med Phys

View full text Add to dashboard Cite

Local noise power spectra (NPS) have been commonly calculated to represent the noise properties of CT imaging systems, but their properties are significantly affected by the utilized calculation schemes. In this study, the effects of varied calculation parameters on the local NPS were analyzed, and practical suggestions were provided regarding the estimation of local NPS for clinical CT scanners. The uniformity module of a Catphan phantom was scanned with a Philips Brilliance 64 slice CT simulator with varied scanning protocols. Images were reconstructed using FBP and iDose4 iterative reconstruction with noise reduction levels 1, 3, and 6. Local NPS were calculated and compared for varied region of interest (ROI) locations and sizes, image background removal methods, and window functions. Additionally, with a predetermined NPS as a ground truth, local NPS calculation accuracy was compared for computer simulated ROIs, varying the aforementioned parameters in addition to ROI number. An analysis of the effects of these varied calculation parameters on the magnitude and shape of the NPS was conducted. The local NPS varied depending on calculation parameters, particularly at low spatial frequencies below ∼0.15 mm−1. For the simulation study, NPS calculation error decreased exponentially as ROI number increased. For the Catphan study the NPS magnitude varied as a function of ROI location, which was better observed when using smaller ROI sizes. The image subtraction method for background removal was the most effective at reducing low‐frequency background noise, and produced similar results no matter which ROI size or window function was used. The PCA background removal method with a Hann window function produced the closest match to image subtraction, with an average percent difference of 17.5%. Image noise should be analyzed locally by calculating the NPS for small ROI sizes. A minimum ROI size is recommended based on the chosen radial bin size and image pixel dimensions. As the ROI size decreases, the NPS becomes more dependent on the choice of background removal method and window function. The image subtraction method is most accurate, but other methods can achieve similar accuracy if certain window functions are applied. All dependencies should be analyzed and taken into account when considering the interpretation of the NPS for task‐based image quality assessment.PACS number(s): 87.57.C‐, 87.57.Q‐

show abstract

Unified SNR analysis of medical imaging systems

Cited by 343 publications

References 59 publications

An adaptive reference/test paradigm: Application to pulsed fluoroscopy perception

An adaptive reference/test paradigm: Application to pulsed fluoroscopy perception

Phantoms and automated system for testing the resolution of ultrasound scanners

Practical considerations for noise power spectra estimation for clinical CT scanners

Contact Info

Product

Resources

About