Evaluation of non-linear adaptive smoothing filter by digital phantom

As a result of the development of multi-slice CT, diagnoses based on three-dimensional reconstruction images and multi-planar reconstruction have spread. For these applications, which require high z-resolution, thin slice imaging is essential. However, because z-resolution is always based on a trade-off with image noise, thin slice imaging is necessarily accompanied by an increase in noise level. To improve the quality of thin slice images, a non-linear adaptive smoothing filter has been developed, and is being widely applied to clinical use. We developed a digital bar pattern phantom for the purpose of evaluating the effect of this filter and attempted evaluation from an addition image of the bar pattern phantom and the image of the water phantom. The effect of this filter was changed in a complex manner by the contrast and spatial frequency of the original image. We have confirmed the reduced effect of image noise in the low frequency component of the image, but decreased contrast or increased quantity of noise in the image of the high frequency component. This result represents the effect of change in the adaptation of this filter. The digital phantom was useful for this evaluation, but to understand the total effect of filtering, much improvement of the shape of the digital phantom is required.     

Fourier block noise reduction: an adaptive filter for reducing Poisson noise in scintigraphic images

OBJECTIVES: Both qualitative and quantitative analysis in nuclear medicine can be undermined by Poisson noise in low-count clinical images. Whilst the conventional smoothing filters are typically used do reduce noise, they also degrade the image structure. Fourier block noise reduction (FBNR) is an adaptive filtering approach, which attempts to reduce image noise and maintain image resolution and structure. METHODS: Although a degree of automated flexibility is possible using conventional stationary pre-filtering, e.g. using a total image count-dependent Metz filter, resolution and contrast is degraded across the image. Adaptive non-stationary filtering has been applied by others in an attempt to maintain structure whilst reducing noise: instead of analysing the whole image, only a subset is used to determine each pixel's correction. Whilst the new software algorithm FBNR shares some common components with other adaptive non-stationary filters, it expressly includes the Poisson noise model within a simple and robust algorithm that can be applied to a diverse range of clinical studies. RESULTS AND CONCLUSIONS: No additional artefacts were seen post-application of FBNR during evaluation using simulated and clinical images. Mean normalised error values indicate FBNR processing is equivalent to obtaining an unprocessed image with at least 2.5 times the number of counts.     

Adaptive filtering for high resolution magnetic resonance images

We present a study of least mean square (LMS) based adaptive filters for high resolution magnetic resonance (MR) images to improve signal-to-noise ratio (SNR) while maintaining sharp edges. Five variations of a new technique that senses the type of noise or the presence of an edge in the filtering window, called adaptive filtering with noise estimation (AFEN) are presented and compared with the basic two-dimensional LMS (TDLMS) algorithm, adaptive filtering with a mean estimator (AFLME), a two-dimensional averaged LMS (TDALMS) algorithm, and a two-dimensional median weighted LMS (TDMLMS) algorithm. Although TDLMS, TDALMS, and TDMLMS filters give better SNR improvement when applied uniformly to an image, they significantly blur edges. The AFLME and AFEN filters both show approximately a factor of 2 SNR improvement with much better retention of edges, with AFEN showing slightly better performance for both SNR and edge sharpness in phantom and in vivo inner ear images.     

Performance evaluation of computed tomography with equivalent resolution images

A method of evaluating the performance of computed tomography (CT) with equivalent resolution images was investigated. Generally, in performance evaluations of CT, the resolution property is measured by the wire method, and the noise property is measured from noise images of a cylindrical water phantom. The signal-to-noise ratio (SNR) is then calculated for the integrated evaluation. Our proposed method enabled perceptual integrated evaluation by using equivalent resolution images created with frequency processing. The frequency-processing factor was calculated from the ratio of the modulation transfer factors of two models of CT, and the image of one of the two was processed by the factor. Because these processed images have resolution equivalent to images of the other CT, the perceptual evaluation with noise images becomes effective. In this investigation, images of a water phantom and a middle contrast resolution phantom were employed. Perceptual comparison of the amount of noise with equivalent resolution images could be performed easily, and effective performance evaluation was achieved. Therefore, our proposed method is useful for noise property and performance evaluation of CT.     

混叠效应对数字化X线影像的影响及其分析

目的 探讨数字X线摄影中混叠效应的产生原因及其对影像质量的影响。方法 使用Kodak CR900系统分别对矩形波测试卡和固定滤线栅成像。测试卡铅条在平行于激光扫描方向和呈45°角的放置方式下分别成像,观察两种影像对分辨率测试卡的显示情况。使滤线栅铅条的方向在平行和垂直于激光扫描方向时分别成像,在显示器上使用不同的放大率对滤线栅影像进行观察,并比较两种影像外观的差异。结果 测试卡影像上,3.93线对/mm以上频率的铅条影无法相互分辨,因此尼奎斯特频率为3.93 LP/mm。但在测试卡45°放置的影像中,频率为4.86 LP/mm的铅条仍能相互分辨。滤线栅铅条平行于成像板的激光扫描方向时,最终影像中显示出明显的条纹影。在显示器上使用不同的放大率观察时,栅条形成的条纹影宽度和方向都发生明显变化。铅条与激光扫描方向垂直时影像中的条纹影显著减弱。结论 CR应用中混叠效应会对数字影像产生不利影响,应根据设备的极限分辨率选择适当的固定滤线栅,栅频率应大于尼奎斯特频率。在摄影中,滤线栅的铅条方向应与成像板的激光扫描方向相垂直。在显示器上观察固定滤线栅摄影影像时,建议使用原始图像的整数倍的放大率。     

Digital radiography of scoliosis with a scanning method: initial evaluation

PURPOSE: To evaluate the radiation dose, image quality, and Cobb angle measurements obtained with a digital scanning method of scoliosis radiography. MATERIALS AND METHODS: Multiple images were reconstructed into one image at a workstation. A low-dose alternative was to use digital pulsed fluoroscopy. Dose measurements were performed with thermoluminescent dosimeters in an Alderson phantom. At the same time, kerma area-product values were recorded. A Monte Carlo dose calculation also was performed. Image quality was evaluated with a contrast-detail phantom and visual grading system. Angle measurements were evaluated with an angle phantom and measurements obtained on patient images. RESULTS: The effective radiation dose was 0.087 mSv for screen-film imaging, 0.16 mSv for digital exposure imaging, and 0.017 mSv for digital fluoroscopy; the corresponding kerma area-product values were 0.43, 0.87, and 0.097 Gy. cm(2), respectively. The image quality of the digital exposure and screen-film images was about equal at visual grading, whereas fluoroscopy had lower image quality. The angle phantom had lower angle values with digital fluoroscopy, although the difference in measured angles was less than 0.5 degrees. The patient images showed no difference in angles. CONCLUSION: The described digital scanning method has acceptable image quality and adequate accuracy in angle measurements. The radiation dose required for digital exposure imaging is higher than that required for screen-film imaging, but that required for digital fluoroscopy is much lower.     

Super‐resolution for multislice diffusion tensor imaging

Diffusion weighted magnetic resonance images are often acquired with single shot multislice imaging sequences, because of their short scanning times and robustness to motion. To minimize noise and acquisition time, images are generally acquired with either anisotropic or isotropic low resolution voxels, which impedes subsequent posterior image processing and visualization. In this article, we propose a super‐resolution method for diffusion weighted imaging that combines anisotropic multislice images to enhance the spatial resolution of diffusion tensor data. Each diffusion weighted image is reconstructed from a set of arbitrarily oriented images with a low through‐plane resolution. The quality of the reconstructed diffusion weighted images was evaluated by diffusion tensor metrics and tractography. Experiments with simulated data, a hardware DTI phantom, as well as in vivo human brain data were conducted. Our results show a significant increase in spatial resolution of the diffusion tensor data while preserving high signal to noise ratio. Magn Reson Med, 2013. © 2012 Wiley Periodicals, Inc.     

Measurement of signal-to-noise ratios in MR images: influence of multichannel coils, parallel imaging, and reconstruction filters

PURPOSE: To evaluate the validity of different approaches to determine the signal-to-noise ratio (SNR) in MRI experiments with multi-element surface coils, parallel imaging, and different reconstruction filters. MATERIALS AND METHODS: Four different approaches of SNR calculation were compared in phantom measurements and in vivo based on: 1) the pixel-by-pixel standard deviation (SD) in multiple repeated acquisitions; 2) the signal statistics in a difference image; and 3) and 4) the statistics in two separate regions of a single image employing either the mean value or the SD of background noise. Different receiver coil systems (with one and eight channels), acquisitions with and without parallel imaging, and five different reconstruction filters were compared. RESULTS: Averaged over all phantom measurements, the deviations from the reference value provided by the multiple-acquisitions method are 2.7% (SD 1.6%) for the difference method, 37.7% (25.9%) for the evaluation of the mean value of background noise, and 34.0% (38.1%) for the evaluation of the SD of background noise. CONCLUSION: The conventionally determined SNR based on separate signal and noise regions in a single image will in general not agree with the true SNR measured in images after the application of certain reconstruction filters, multichannel reconstruction, or parallel imaging.     

Xenon versus ceramics: a comparison of two CT X-ray detector systems.

PURPOSE: The purpose of this work was to compare image quality in phantom and patient CT scans acquired by xenon and ceramic CT detector systems. METHOD: High and low contrast resolution and image noise were determined with a standard CT phantom for both detector systems. In patient CT images, the effect on image noise was measured in anatomical regions of interest in the head, lumbar spine, chest, and abdomen. RESULTS: In phantom studies, image noise was significantly lower using ceramic versus xenon detectors. Also, in images of the head and lumbar spine, the signal-to-noise ratio was significantly higher with ceramic than with xenon detectors. In chest scans, ceramic significantly reduced beam-hardening artifacts caused by the thoracic spine. However, in abdominal images, the signal-to-noise ratio was not significantly different between ceramic and xenon detector systems. CONCLUSION: For reduced image noise in CT images of the head, lumbar spine, and chest and high resolution CT, ceramic detector systems appear to be superior to xenon detector systems.     

Spatial resolution measurement for iterative reconstruction by use of image-averaging techniques in computed tomography

The purpose of our study was to investigate the validity of a spatial resolution measuring method that uses a combination of a bar-pattern phantom and an image-averaging technique, and to evaluate the spatial resolution property of iterative reconstruction (IR) images with middle-contrast (50 HU) objects. We used computed tomography (CT) images of the bar-pattern phantom reconstructed by the IR technology Adaptive Iterative Dose Reduction 3D (AIDR 3D), which was installed in the multidetector CT system Aquilion ONE (Toshiba Medical Systems, Otawara, Japan). The contrast of the bar-pattern image was set to 50 HU, which is considered to be a middle contrast that requires higher spatial resolution clinically. We employed an image-averaging technique to eliminate the influence of image noise, and we obtained averaged images of the bar-pattern phantom with sufficiently low noise. Modulation transfer functions (MTFs) were measured from the images. The conventional wire method was also used for comparison; in this method, AIDR 3D showed MTF values equivalent to those of filtered back projection. For the middle-contrast condition, the results showed that the MTF of AIDR 3D decreased with the strength of IR processing. Further, the MTF of AIDR 3D decreased with dose reduction. The image-averaging technique used was effective for correct evaluation of the spatial resolution for middle-contrast objects in IR images. The results obtained by our method clarified that the resolution preservation of AIDR 3D was not sufficient for middle-contrast objects.     

Image reconstruction in SNR units: a general method for SNR measurement.

The method for phased array image reconstruction of uniform noise images may be used in conjunction with proper image scaling as a means of reconstructing images directly in SNR units. This facilitates accurate and precise SNR measurement on a per pixel basis. This method is applicable to root-sum-of-squares magnitude combining, B(1)-weighted combining, and parallel imaging such as SENSE. A procedure for image reconstruction and scaling is presented, and the method for SNR measurement is validated with phantom data. Alternative methods that rely on noise only regions are not appropriate for parallel imaging where the noise level is highly variable across the field-of-view. The purpose of this article is to provide a nuts and bolts procedure for calculating scale factors used for reconstructing images directly in SNR units. The procedure includes scaling for noise equivalent bandwidth of digital receivers, FFTs and associated window functions (raw data filters), and array combining.     

Two-dimensional filtering of SPECT images using the Metz and Wiener filters

Two-dimensional filtering, both before and after reconstruction, has been applied to the processing of single photon emission computerized tomographic (SPECT) images. The filters investigated were the count-dependent Metz filter and Wiener filter, both of which automatically adapt to the image being processed. Using a SPECT phantom, with images reconstructed with these filters rather than the ramp, we observed a statistically significant increase (p less than 0.05) in the image contrast for solid Plexiglas spheres, and significant decrease (p less than 0.05) in the percent fractional standard deviation of counts in a region of uniform activity. The adaptability of these filters is demonstrated by a comparison of SPECT acquisitions of the phantom at two different count levels. An example of their application to clinical studies is presented. We conclude that two-dimensional digital image restoration with these techniques can produce a significant increase in SPECT image quality, with a small cost in processing time when these techniques are implemented on an array processor.     

Functional image data acquisition and processing

In the production of functional images certain hardware and software considerations are necessary for the rapid and accurate determination of kinetic parameters. The advent of the digital scintillation camera has made available increased accuracy of quantitation and ease of image handling, although its integrated computer system may not be optimal for program development presently. To reduce the deleterious effects of Poisson noise on parameter estimation, the single or multiple application of easily implemented smoothing operators in space and time is recommended as a first step in image processing. The properties of these operators are conveniently expressed in terms of their variance. Following smoothing, count or variance thresholding is performed to reduce computer processing time and eliminate extraneous background from functional images. Time-activity curves can be fit by a variety of mathematical functions, the most useful of which is probably the finite Fourier series. In a simulated gated blood-pool study of the left ventricle, with and without an aneurysm, it is found that increased smoothing of the original image data results in more accurate parameter determinations, to the extent that small regions of dissimilar temporal behavior are not obliterated.     

A method to optimize the processing algorithm of a computed radiography system for chest radiography

A test methodology using an anthropomorphic-equivalent chest phantom is described for the optimization of the Agfa computed radiography "MUSICA" processing algorithm for chest radiography. The contrast-to-noise ratio (CNR) in the lung, heart and diaphragm regions of the phantom, and the "system modulation transfer function" (sMTF) in the lung region, were measured using test tools embedded in the phantom. Using these parameters the MUSICA processing algorithm was optimized with respect to low-contrast detectability and spatial resolution. Two optimum "MUSICA parameter sets" were derived respectively for maximizing the CNR and sMTF in each region of the phantom. Further work is required to find the relative importance of low-contrast detectability and spatial resolution in chest images, from which the definitive optimum MUSICA parameter set can then be derived. Prior to this further work, a compromised optimum MUSICA parameter set was applied to a range of clinical images. A group of experienced image evaluators scored these images alongside images produced from the same radiographs using the MUSICA parameter set in clinical use at the time. The compromised optimum MUSICA parameter set was shown to produce measurably better images.     

A digital fluoroscopic imaging system for verification during external beam radiotherapy

A digital fluoroscopic (DF) imaging system has been constructed to obtain portal images for verification during external beam radiotherapy. The imaging device consists of a fluorescent screen viewed by a highly sensitive video camera through a mirror. The video signal is digitized and processed by an image processor which is linked on-line with a host microcomputer. The image quality of the DF system was compared with that of film for portal images of the Burger phantom and the Alderson anthropomorphic phantom using 10 MV X-rays. Contrast resolution of the DF image integrated for 8.5 sec. was superior to the film resolution, while spatial resolution was slightly inferior. The DF image of the Alderson phantom processed by the adaptive histogram equalization was better in showing anatomical landmarks than the film portal image. The DF image integrated for 1 sec. which is used for movie mode can show patient movement during treatment.     

Weighted-averaging multi-planar reconstruction method for multi-detector row computed tomography

Development of multi-detector row computed tomography (MDCT) has enabled three-dimensions (3D) scanning with minute voxels. Minute voxels improve spatial resolution of CT images. At the same time, however, they increase image noise. Multi-planar reconstruction (MPR) is one of effective 3D-image processing techniques. The conventional MPR technique can adjust slice thickness of MPR images. When a thick slice is used, the image noise is decreased. In this case, however, spatial resolution is deteriorated. In order to deal with this trade-off problem, we have developed the weighted-averaging multi-planar reconstruction (W-MPR) technique to control the balance between the spatial resolution and noise. The weighted-average is determined by the Gaussian-type weighting function. In this study, we compared the performance of W-MPR with that of conventional simple-addition-averaging MPR. As a result, we could confirm that W-MPR can decrease the image noise without significant deterioration of spatial resolution. W-MPR can adjust freely the weight for each slice by changing the shape of the weighting function. Therefore, W-MPR can allow us to select a proper balance of spatial resolution and noise and at the same time produce suitable MPR images for observation of targeted anatomical structures.     

Usefulness of noise adaptive non-linear Gaussian filter in FDG-PET study

OBJECTIVE: In positron emission tomography (PET) studies, shortening transmission (TR) scan time can improve patient comfort and increase scanner throughput. However, PET images from short TR scans may be degraded due to the statistical noise included in the TR image. The purpose of this study was to apply non-linear Gaussian (NLG) and noise adaptive NLG (ANLG) filters to TR images, and to evaluate the extent of noise reduction by the ANLG filter in comparison with that by the NLG filter using phantom and clinical studies. METHODS: In phantom studies, pool phantoms of various diameters and injected doses of 2-deoxy-2-[18F]fluoro-D-glucose (FDG) were used and the coefficients of variation (CVs) of the counts in the TR images processed with the NLG and ANLG filters were compared. In clinical studies, two normal volunteers and 13 patients with tumors were studied. In volunteer studies, the CV values in the liver were compared. In patient studies, the standardized uptake values (SUVs) of tumors in the emission images were obtained after processing the TR images using the NLG and ANLG filters. RESULTS: In phantom studies, the CV values in the TR images processed with the ANLG filter were smaller than those in the images processed with the NLG filter. When using the ANLG filter, their dependency on the phantom size, injected dose of FDG and TR scan time was smaller than when using the NLG filter. In volunteer studies, the CV values in the images processed with the ANLG filter were smaller than those in the images processed with the NLG filter, and were almost constant regardless of the TR scan time. In patient studies, there was an excellent correlation between the SUVs obtained from the images with a TR scan time of 7 min processed with the NLG filter (x) and those obtained from the images with a TR scan time of 4 min processed with the ANLG filter (y) (r = 0.995, y = 1.034x - 0.075). CONCLUSIONS: Our results suggest that the ANLG filter is effective and useful for noise reduction in TR images and shortening TR scan time while maintaining the quantitative accuracy of FDG-PET studies.     

Adaptive statistical iterative reconstruction for computed tomography of the spine

IntroductionThe utility of evaluating a sagittal view of CT of the spine is well-known. In many clinical cases, the sagittal view includes noise generated from surrounding objects and may degrade the image quality. Iterative reconstruction (IR) techniques are useful for noise reduction; however, they can reduce spatial resolution. The aim of this study was to evaluate the effectiveness of the adaptive statistical iterative reconstruction (ASiR) for generating sagittal CT images of the spine when compared to filtered back projection (FBP).MethodsThe image quality of clinical images from 25 patients were subjectively assessed. Three radiologists rated spatial resolution, image noise, and overall image quality using a five-point scale. For objective assessment, z-direction modulation transfer function (z-MTF) was measured using a custom-made phantom. Additionally, z-axis noise power spectrum (z-NPS) was measured using a water phantom. An improved adaptive statistical iterative reconstruction algorithm called ASiR-V was used in this study. Blending levels were 50%, and 100% (ASiR-V50, ASiR-V100, respectively).ResultsFor subjective assessments, images using ASiR-V100 were determined to have the best overall image quality, despite having received the worst score in the assessment of spatial resolution. For objective assessments, the image using ASiR-V50 and ASiR-V100 curves were slightly degraded in terms of low contrast z-MTF when compared to FBP.ConclusionASiR-V was effective to improve the image quality when compared with FBP when reviewing sagittal reformats of the spine.Implications for practiceThis study suggests that high resolution is not the only thing that is key when reviewing sagittal CT spinal reformats. Such images should be provided as part of routine CT spine protocols, where available.     

Evaluation of reconstruction techniques for lung single photon emission tomography: a Monte Carlo study

BACKGROUND: In studies of the distribution of lung function, the image quality of lung single photon emission computed tomography (SPECT) is important and one factor influencing it is the reconstruction algorithm. AIM: To systematically evaluate ordered subsets expectation maximization (OSEM) and compare it with filtered back-projection (FBP) for lung SPECT with Tc. METHODS: The evaluation of the number of iterations used in OSEM was based on the image quality parameter contrast. The comparison between OSEM and FBP was based on trade-off plots between statistical noise and spatial resolution for different filter parameters, collimators and count-levels. A Monte Carlo technique was used to simulate SPECT studies of a digital thorax phantom containing two sets of activity: one with a homogeneous activity distribution within the lungs and the other with superposed high- and low-activity objects. Statistical noise in the reconstructed images was calculated as the coefficient of variation (CV) and spatial resolution as full width at half-maximum (FWHM). RESULTS: For the configuration studied, the OSEM reconstruction in combination with post-filtering should be used in lung SPECT studies with at least 60 MLEM equivalent iterations. Compared to FBP the spatial resolution was improved by about 1 mm. For a constant level of CV, a four-fold increase in count level resulted in an increased resolution of about 2 mm. Spatial resolution and cut-off frequency depends on what value of noise in the image is acceptable also increased by using a low-energy, high-resolution collimator for CV values above 3%. The choice of noise-reducing filter and cut-off frequency depends on what value of noise in the image is acceptable.     

Performance evaluation for CT-AEC(CT automatic exposure control)systems

Although many current CT scanners incorporate CT-AEC, performance evaluation is not standardized. This study evaluates the performance of the latest CT-AEC of each manufacturer with the aim of establishing a standard CT-AEC performance evaluation method. The design of the phantoms was based upon the operation characteristics of different CT-AECs. A cone, an ellipse, a variable-shaped ellipse, stepped phantoms, and their analysis software were devised and carried out the field test. The targets were LightSpeed VCT 64 with 2D and 3D Auto mA(GE), Aquilion 64M with Real-EC and Volume-EC(Toshiba), Sensation 64 with CARE Dose and CARE Dose 4D(Siemens), and Bulliance 16P with Dose Right(Philips). Data was acquired while varying the typical abdominal CT(with CT-AEC)scanning conditions (120 kV, 5 mm slice, standard function for abdomen, scanning range 200 mm). The acquired images were converted to the DICOM format and image noise(SD) was calculated using dedicated software. All 4 CT-AECs reduced exposure dose. For GE and Toshiba, image noise was constant and met the target. For Siemens, noise was independent of phantom shape but varied uniformly with phantom size. For Philips, noise varied with phantom size and shape, and variation degree depended on phantom thickness in scanogram direction. The results reflect the basic concept and performance characteristics of the methods. Standardization of CT-AEC performance evaluation is possible using these phantoms.