Determination of point spread function in computed tomography accompanied with verification

Abstract

A method for verifying the point spread function (PSF) measured by computed tomography has been previously reported [Med. Phys. 33, 2757-2764 (2006)]; however, this additional PSF verification following measurement is laborious. In the present study, the previously described verification method was expanded to PSF determination. First, an image was obtained by scanning a phantom. The image was then two-dimensionally deconvolved with the object function corresponding to the phantom structure, thus allowing the PSF to be obtained. Deconvolution is implemented simply by division of spatial frequencies (corresponding to inverse filtering), in which two parameters are used as adjustable ones. Second, an image was simulated by convolving the object function with the obtained PSF, and the simulated image was compared to the above-measured image of the phantom. The difference indicates the inaccuracy of the PSF obtained by deconvolution. As a criterion for evaluating the difference, the authors define the mean normalized standard deviation (SD) in the difference between simulated and measured images. The above two parameters for deconvolution can be adjusted by referring to the subsequent mean normalized SD (i.e., the PSF is determined so that the mean normalized SD is decreased). In this article, the parameters were varied in a fixed range with a constant increment to find the optimal parameter setting that minimizes the mean normalized SD. Using this method, PSF measurements were performed for various types of image reconstruction kernels (21 types) in four kinds of scanners. For the 16 types of kernels, the mean normalized SDs were less than 2.5%, indicating the accuracy of the determined PSFs. For the other five kernels, the mean normalized SDs ranged from 3.7% to 4.8%. This was because of a large amount of noise in the measured images, and the obtained PSFs would essentially be accurate. The method effectively determines the PSF, with an accompanying verification, after one scanning of a phantom.

Generation of realistic virtual nodules based on three‐dimensional spatial resolution in lung computed tomography: A pilot phantom study

“…A two‐dimensional (2D) point spread function (PSF) in the x–y scanning plane and a slice sensitivity profile (SSP) in the z direction perpendicular to the scanning plane were measured for each scanner. The measurements of PSFs for reconstruction kernels of FC52 (Scanner A), FC14 (Scanner B), and Bf44d (Scanner C) were performed with a previously proposed method that determined the PSF and verified its accuracy . The measurements of SSPs for a slice thickness of 1 mm in all scanners were obtained using the micro coin phantom (Kyoto Kagaku, Kyoto, Japan).…”

Accuracy of lung nodule density on HRCT: analysis by PSF‐based image simulation

“…The PSFs were measured in the scanner by a recently proposed method that can be used to determine the 2D PSF in the scan plane with one scan of the phantom described above (Materials & Methods Section D), accompanied by verification 16 , 17 . Another section of the phantom including five cylindrical objects (diameter of 2, 3, 5, 7, and 10 mm) was used.…”

“…1b). This simulated image has the advantage of having an exact dependency on the characteristics of the spatial resolution of the CT system 14 , 17 . However, the image has a fine interval of the discrete data; therefore, we propose an additional resampling to obtain the clinical CT image such as the image indicated in (Fig.…”

A method for evaluating the performance of computer-aided detection of pulmonary nodules in lung cancer CT screening: detection limit for nodule size and density

“…When scanning a point-source phantom, common methods, such as using a thin wire or a microbead, have known difficulties in obtaining an accurate PSF. 27 We submit that the 2D PSF measurement method accompanied by verification, 22,23 which was used in the present study, yields sound results for validation.…”

“…The PSF was determined by scanning a high-contrast CT test phantom (MHT-type; Kyoto Kagaku Co. Ltd, Kyoto, Japan); this method includes a means of verifying the obtained PSF and is therefore considered to give accurate values. 22,23 The SSP was measured with the use of a Gold Disk Delta phantom (Kyoto Kagaku Co. Ltd), comprising a gold disc 50-mm thick and 1.0 mm in diameter placed in a tissue-equivalent material (acrylic).…”