Do-Il Kim

Reduction of a grid moiré pattern by integrating a carbon-interspaced high precision x-ray grid with a digital radiographic detector.

The stationary grid commonly used with a digital x-ray detector causes a moiré interference pattern due to the inadequate sampling of the grid shadows by the detector pixels. There are limitations with the previous methods used to remove the moiré such as imperfect electromagnetic interference shielding and the loss of image information. A new method is proposed for removing the moiré pattern by integrating a carbon-interspaced high precision x-ray grid with high grid line uniformity with the detector for frequency matching. The grid was aligned to the detector by translating and rotating the x-ray grid with respect to the detector using microcontrolled alignment mechanism. The gap between the grid and the detector surface was adjusted with micrometer precision to precisely match the projected grid line pitch to the detector pixel pitch. Considering the magnification of the grid shadows on the detector plane, the grids were manufactured such that the grid line frequency was slightly higher than the detector sampling frequency. This study examined the factors that affect the moiré pattern, particularly the line frequency and displacement. The frequency of the moiré pattern was found to be sensitive to the angular displacement of the grid with respect to the detector while the horizontal translation alters the phase but not the moiré frequency. The frequency of the moiré pattern also decreased with decreasing difference in frequency between the grid and the detector, and a moiré-free image was produced after complete matching for a given source to detector distance. The image quality factors including the contrast, signal-to-noise ratio and uniformity in the images with and without the moiré pattern were investigated.

Performance evaluation of carbon-interspaced antiscatter grids in mammography: Empirical formula and Monte Carlo simulation studies

Anti-scatter grids show various performances depending on the complicated design parameters. However, it takes long time and cost to fabricate grids and perform experiments to evaluate the performance of grids. Therefore, to replace actual measurements, we performed Monte Carlo simulation study and obtained Tp (transmission of primary radiation) and Tt (transmission of total radiation) values which represent grid performance. Beam quality check was done to validate our simulation results. Although the simulation lowers the cost, it is still time consuming work. To solve those problems, empirical formulae for Tp and Tt were derived by analyzing the results of the simulation. Design parameters for grids were setup in order to derive the empirical formulae. The derived empirical formulae helped us estimate Tp and Tt values of an arbitrary grid without a complicated experiment or a tedious simulation, and easily obtain contrast improvement factor (CIF) and Bucky factor (BF) which are commonly used to evaluate the performance of a grid. The design parameters of a grid for the minimum patients dose can be proposed with maintaining the image quality within the designable range using the derived empirical formulae.

Image Quality Analysis of an aSi:H/CsI(Tl) Flat-Panel Based Digital Radiography System Using a Chest Phantom

In digital radiography system, the diagnostic utility of image quality mainly depends on processing technique of image data. The aim of this work is to quantitatively evaluate image quality of our aSi:H/CsI(Tl) flat-panel based digital radiography system using a chest phantom. The study on the human visual property and the detector response to four standard radiation qualities (from RQA 3 to RQA 9) was performed. The high and low contrast resolution was assessed with two commercially available digital systems based on the same detector. The resulting signal-to-noise ratios (SNRs) of the nine low contrast objects in the lung, heart and subdia-phragm regionns, and relative spatial resolution in the lung field were comparable with those of other systems. Low contrast objects were automatically detected and SNRs was computed by our image analysis algorithm. The quantitative analysis procedure of image quality designed in this work removes observer’s subjectivity. In addition, It can be easily applied for the detectability measurement of low contrast object in other digital modalities. by our Image analysis algorithm

SU‐FF‐I‐23: An Iterative Method for Flat‐Field Correction of Digital Radiography with Arbitrary Detector Position

Purpose: To investigated the effect of X‐ray tube positions with respect to detector on the non‐uniformity correction, and propose a method to reduce the effect using a new algorithm with computer simulation.Method and Materials: Gain images that represented the reference image used to execute the flat field correction were taken in two SIDs (Source to Image‐receptor Distance). Pixel values at second SID was calculated using the pixel values at first SID, the assumed gain coefficient, and the formula based on the solid angle of each detector pixel facing to the x‐ray source. Gain coefficient was adjusted using the difference between calculated and real pixel values at second SID. Calculation was repeated with new gain coefficient until the gain coefficient was converged into prescribed range. Flat field correction could be performed using acquired gain image. Non‐uniformity of blank x‐ray images taken with the detector tilted by 0 to 45 degrees was corrected and five ROIs across the image were defined and analyzed. Results: With a blank image obtained with the detector tilting angle of 45 degrees the lowest ROI mean value was 53% less than the highest ROI mean value when usual non‐uniformity correction was performed. When the proposed was used for the flat field correction, however, the lowest ROI mean value was only 7% less than the highest ROI mean value, and standard deviations of pixel values in the ROIs were reduce to 10% of the cases of usual flat field correction. Conclusion: Because of the characteristic of usual flat field correction, non‐uniformity in detector that was not aligned to the X‐ray source considerably increased. We also calculated the gain coefficient and performed the flat field correction with the iterative method in the tilted or arbitrary detector position. The proposed algorithm gave a satisfactory uniformity.

An iterative method for flat-field correction of digital radiography when detector is at any position

For non-uniformity correction a flat field x-ray image is needed, and to obtain it the center of detector is usually aligned with the focal spot of the x-ray tube, which is conserved when examining patients to preserve the flat field. In some of radiographic techniques, however, it is necessary to move the x-ray tube off the center position of detector or tilt the detector. We investigated the effect of X-ray tube positions with respect to detector on the non-uniformity correction, and propose a method to reduce the effect using a new algorithm with computer simulation. Gain images were taken in two SIDs. Pixel values at second SID was calculated using the pixel values at first SID, gain coefficient that represents pixels own unique radiation sensitivity characteristics and the formula based on the solid angle of each detector pixel facing to the x-ray source. Gain coefficient was adjusted using the difference between calculated and real pixel values. Calculation was repeated with new gain coefficient until the gain coefficient was converged into prescribed range. Non-uniformity of blank x-ray images taken with the detector tilted by 0 to 45 degrees was corrected and five ROIs across the image were defined and analyzed. When the proposed algorithm was used for the flat field correction standard deviations of pixel values in the ROIs were reduce to 10% of the cases of usual flat field correction.