Perception of temporally filtered X-ray fluoroscopy images

Assessment of temporal resolution and detectability of moving objects in CT: A task-based image quality study

The metrics used for assessing image quality in computed tomography (CT) do not integrate the influence of temporal resolution. A shortcoming in the assessment of image quality for imaging protocols where motion blur can therefore occur. We developed a method to calculate the temporal resolution of standard CT protocols and introduced a specific spatiotemporal formulation of the non-prewhitening with eye filter (NPWE) model observer to assess the detectability of moving objects as a function of their speed. We scanned a cubic water phantom with a plexiglass cylindrical insert (120 HU) using a large panel of acquisition parameters (rotation times, pitch factors and collimation widths) on two systems (GE Revolution Apex and Siemens SOMATOM Force) to determine the in-plane task-based transfer functions (TTF) and noise power spectra (NPS). The phantom set in a uniform rectilinear motion in the transverse plane allowed the temporal modulation transfer function (MTF) calculation. The temporal MTF appropriately compared the temporal resolution of the various acquisition protocols. The longitudinal TTF was measured using a thin tungsten wire. The detectability index showed the advantage of applying high rotation speed, wide collimations and high pitch for object detection in the presence of motion. No counterpart to the increase in these three parameters was found in the in-plane and longitudinal image quality.

Adaptation of a channelized Hotelling observer model to accommodate anthropomorphic backgrounds and moving test objects in X-ray angiography

BACKGROUND

Prior implementations of the channelized Hotelling observer (CHO) model have succeeded in assessing the performance of X-ray angiography systems under a variety of imaging conditions. However, often times these conditions do not resemble those present in routine clinical imaging scenarios, such as having complex anthropomorphic backgrounds in conjunction with moving test objects.

PURPOSE

This work builds up on prior established CHO methods and introduces a new approach to switch from the already established "multiple-sample" CHO implementation to a "single-sample" technique. The proposed implementation enables the inclusion of moving test objects upon nonuniform backgrounds by allowing only a single sample to represent the test object present condition that is to be used within the statistical test to estimate the detectability index.

METHODS

To assess the proposed method, two image data sets were acquired with a clinical X-ray angiography system. The first set consisted of a uniform background in combination with static test objects while the second consisted of an anthropomorphic chest phantom in conjunction with moving test objects. The first set was used to validate the proposed approach against the multiple-sample method while the second was used to assess the feasibility of the proposed method under a variety of imaging conditions, including seven object sizes and seven detector target dose (DTD) levels.

RESULTS

For the uniform background data set, considering all detectability indices greater or equal than 1, the ratio between the detectability indices of the proposed single-sample approach versus the multiple-sample method was 0.997 ± 0.056 (range 0.884-1.159). The average single-direction width of the 95% confidence intervals (CIs) of the detectability index estimates for the multiple-sample method was 0.38 ± 0.43 (range 0.03-2.20). For the single-sample approach, the average width was 2.52 ± 0.63 (range 1.11-5.44). For the anthropomorphic background image set, the results were consistent with classical quantum-limited signal-to-noise ratio (SNR) theory. The magnitude of the detectability indices varied predictably with changes in both object size and DTD, with the highest value associated with the highest dose and the largest object size. Additionally, the proposed method was able to capture differences in the imaging performance for a given test object across the field of view, which was associated with the attenuation levels provided by different features of the anthropomorphic background.

CONCLUSIONS

A new single-sample variant of the CHO model to assess the performance of X-ray angiography imaging systems is proposed. The new approach is consistent with quantum-limited image quality theory and with a standard implementation of the CHO model. The proposed method enables the assessment of moving test objects in combination with complex, nonuniform image backgrounds, thereby opening the possibility to assess imaging conditions which more closely resemble those used in clinical care.

Real-time X-ray imaging of mouse cerebral microvessels in vivo using a pixel temporal averaging method

Rodents are used extensively as animal models for the preclinical investigation of microvascular-related diseases. However, motion artifacts in currently available imaging methods preclude real-time observation of microvessels in vivo. In this paper, a pixel temporal averaging (PTA) method that enables real-time imaging of microvessels in the mouse brain in vivo is described. Experiments using live mice demonstrated that PTA efficiently eliminated motion artifacts and random noise, resulting in significant improvements in contrast-to-noise ratio. The time needed for image reconstruction using PTA with a normal computer was 250 ms, highlighting the capability of the PTA method for real-time angiography. In addition, experiments with less than one-quarter of photon flux in conventional angiography verified that motion artifacts and random noise were suppressed and microvessels were successfully identified using PTA, whereas conventional temporal subtraction and averaging methods were ineffective. Experiments performed with an X-ray tube verified that the PTA method could also be successfully applied to microvessel imaging of the mouse brain using a laboratory X-ray source. In conclusion, the proposed PTA method may facilitate the real-time investigation of cerebral microvascular-related diseases using small animal models.

A novel method to assess the spatiotemporal image quality in fluoroscopy

Objectives. The planar formulation of the noise equivalent quanta (NEQ) and detective quantum efficiency (DQE) used to assess the image quality of projection images does not deal with the influence of temporal resolution on signal blurring and image noise. These metrics require correction factors based on temporal resolution when used for dynamic imaging systems such as fluoroscopy. Additionally, the standard NEQ and detector DQE are determined on pre-processed images in scatter-free conditions for effective energies produced by additional aluminium or copper filters that are not representative of clinical fluoroscopic procedures. In this work, we developed a method to measure 'frame NEQ' and 'frame system DQE' which include the temporal frequency bandwidth and consider the anti-scatter grid, the detector and the image processing procedures for beam qualities with scatter fractions representative of clinical use.Approach. We used a solid water phantom to simulate a patient and a thin copper disc to measure the spatial resolution. The copper disc, set in uniform rectilinear motion in the image plane, assessed the temporal resolution. These new metrics were tested on two fluoroscopy systems, a C-arm and a floor-mounted cardiology, for multiple parameters: phantom thicknesses from 5 to 20 cm, frame rates from 3 to 30 fps, spatial and temporal image processing of different weights.Main results.The frame NEQ correctly described the image quality for different scatter conditions, temporal resolutions and image processing techniques. The frame system DQE varied between 0.38 and 0.65 within the different beam and scatter conditions, and correctly mitigated the influence of spatial and temporal image processing.Significance.This study introduces and validates an unbiased formulation of in-plane NEQ and system DQE to assess the spatiotemporal image quality of fluoroscopy systems.

Task-specific efficient channel selection and bias management for Gabor function channelized Hotelling observer model for the assessment of x-ray angiography system performance

PURPOSE

Channelized Hotelling observer (CHO) models have been implemented to assess imaging performance in x-ray angiography systems. While current methods are appropriate for assessing unprocessed images of moving test objects upon uniform-exposure backgrounds, they are inadequate for assessing conditions which more appropriately mimic clinical imaging conditions including the combination of moving test objects, complex anthropomorphic backgrounds, and image processing. In support of this broad goal, the purpose of this work was to develop theory and methods to automatically select a subset of task-specific efficient Gabor channels from a task-generic Gabor channel base set. Also, previously described theory and methods to manage detectability index (d') bias due to nonrandom temporal variations in image electronic noise will be revisited herein.

METHODS

Starting with a base set of 96 Gabor channels, backward elimination of channels was used to automatically identify an "efficient" channel subset which reduced the number of channels retained in the subset while maintaining the magnitude of the d' estimate. The concept of a pixelwise Hotelling observer (PHO) model was introduced and similarly implemented to assess the performance of the efficient-channel CHO model. Bias in d' estimates arising from temporally variable nonstationary noise was modeled as a bivariate probability density function for normal distributions, where one variable corresponds to the signal from the test object and the other variable corresponds to the signal from temporally variable nonstationary noise. Theory and methods were tested on uniform-exposure unprocessed angiography images with detector target dose (DTD) of 6, 18, and 120 nGy containing static disk-shaped test objects with diameter in the range of 0.5 to 4 mm.

RESULTS

Considering all DTD levels and test object sizes, the proposed method reduced the number of Gabor channels in the final subset by 63-82% compared to the original 96 Gabor channel base set, while maintaining a mean relative performance ( ( d CHO ' / d PHO ' ) × 100 % ) of 95%  ±  4% with respect to the reference PHO model. Experimental results demonstrated that the bivariate approach to account for bias due to temporally variable nonstationary noise resulted in improved correlation between the CHO and PHO models as compared to a previously proposed univariate approach.

CONCLUSIONS

Computationally efficient backward elimination can be used to select an efficient subset of Gabor channels from an initial channel base set without substantially compromising the magnitude of the d' estimate. Bias due to temporally variable nonstationary noise can be modeled through a bivariate approach leading to an improved unbiased estimate of d'.

Task-based evaluation of a 4D MAP-RBI-EM image reconstruction method for gated myocardial perfusion SPECT using a human observer study

We evaluated the performance of a new 4D image reconstruction method for improved 4D gated myocardial perfusion (MP) SPECT using a task-based human observer study. We used a realistic 4D NURBS-based Cardiac-Torso (NCAT) phantom that models cardiac beating motion. Half of the population was normal; the other half had a regional hypokinetic wall motion abnormality. Noise-free and noisy projection data with 16 gates/cardiac cycle were generated using an analytical projector that included the effects of attenuation, collimator-detector response, and scatter (ADS), and were reconstructed using the 3D FBP without and 3D OS-EM with ADS corrections followed by different cut-off frequencies of a 4D linear post-filter. A 4D iterative maximum a posteriori rescaled-block (MAP-RBI)-EM image reconstruction method with ADS corrections was also used to reconstruct the projection data using various values of the weighting factor for its prior. The trade-offs between bias and noise were represented by the normalized mean squared error (NMSE) and averaged normalized standard deviation (NSDav), respectively. They were used to select reasonable ranges of the reconstructed images for use in a human observer study. The observers were trained with the simulated cine images and were instructed to rate their confidence on the absence or presence of a motion defect on a continuous scale. We then applied receiver operating characteristic (ROC) analysis and used the area under the ROC curve (AUC) index. The results showed that significant differences in detection performance among the different NMSE-NSDav combinations were found and the optimal trade-off from optimized reconstruction parameters corresponded to a maximum AUC value. The 4D MAP-RBI-EM with ADS correction, which had the best trade-off among the tested reconstruction methods, also had the highest AUC value, resulting in significantly better human observer detection performance when detecting regional myocardial wall motion abnormality. We concluded that the NMSE-NSDav trade-off was shown to agree with observer performance for the detection task of the regional motion abnormality, and the optimized 4D MAP-RBI-EM method with ADS corrections provides significant improvement compared to 3D FBP and 3D OS-EM with ADS corrections in detecting regional myocardial wall motion abnormali in 4D gated MP SPECT.

A motion-compensated image filter for low-dose fluoroscopy in a real-time tumor-tracking radiotherapy system

In the real-time tumor-tracking radiotherapy system, a surrogate fiducial marker inserted in or near the tumor is detected by fluoroscopy to realize respiratory-gated radiotherapy. The imaging dose caused by fluoroscopy should be minimized. In this work, an image processing technique is proposed for tracing a moving marker in low-dose imaging. The proposed tracking technique is a combination of a motion-compensated recursive filter and template pattern matching. The proposed image filter can reduce motion artifacts resulting from the recursive process based on the determination of the region of interest for the next frame according to the current marker position in the fluoroscopic images. The effectiveness of the proposed technique and the expected clinical benefit were examined by phantom experimental studies with actual tumor trajectories generated from clinical patient data. It was demonstrated that the marker motion could be traced in low-dose imaging by applying the proposed algorithm with acceptable registration error and high pattern recognition score in all trajectories, although some trajectories were not able to be tracked with the conventional spatial filters or without image filters. The positional accuracy is expected to be kept within ±2 mm. The total computation time required to determine the marker position is a few milliseconds. The proposed image processing technique is applicable for imaging dose reduction.

Implementation of a channelized Hotelling observer model to assess image quality of x-ray angiography systems

Evaluation of flat-panel angiography equipment through conventional image quality metrics is limited by the scope of standard spatial-domain image quality metric(s), such as contrast-to-noise ratio and spatial resolution, or by restricted access to appropriate data to calculate Fourier domain measurements, such as modulation transfer function, noise power spectrum, and detective quantum efficiency. Observer models have been shown capable of overcoming these limitations and are able to comprehensively evaluate medical-imaging systems. We present a spatial domain-based channelized Hotelling observer model to calculate the detectability index (DI) of our different sized disks and compare the performance of different imaging conditions and angiography systems. When appropriate, changes in DIs were compared to expectations based on the classical Rose model of signal detection to assess linearity of the model with quantum signal-to-noise ratio (SNR) theory. For these experiments, the estimated uncertainty of the DIs was less than 3%, allowing for precise comparison of imaging systems or conditions. For most experimental variables, DI changes were linear with expectations based on quantum SNR theory. DIs calculated for the smallest objects demonstrated nonlinearity with quantum SNR theory due to system blur. Two angiography systems with different detector element sizes were shown to perform similarly across the majority of the detection tasks.

Overcoming x-ray tube small focal spot output limitations for high-resolution region of interest imaging

We investigate methods to increase x-ray tube output to enable improved quantum image quality with a higher generalized-NEQ (GNEQ) while maintaining a small focal-spot size for the new high-resolution Micro-angiographic Fluoroscope (MAF) Region of Interest (ROI) imaging system. Rather than using a larger focal spot to increase tube-loading capacity with degraded resolution, we evaluated separately or in combination three methods to increase tube output: 1) reducing the anode angle and lengthening the filament to maintain a constant effective small focal-spot size, 2) using the standard medium focal spot viewed from a direction on the anode side of the field and 3) increasing the frame rate (frames/second) in combination with temporal filter. The GNEQ was compared for the MAF for the small focal-spot at the central axis, and for the medium focal-spot with a higher output on the anode side as well as for the small focal spot with different temporal recursive filtering weights. A net output increase of about 4.0 times could be achieved with a 2-degree anode angle (without the added filtration) and a 4 times longer filament compared to that of the standard 8-degree target. The GNEQ was also increased for the medium focal-spot due to its higher output capacity and for the temporally filtered higher frame rate. Thus higher tube output, while maintaining a small effective focal-spot, should be achievable using one or more of the three methods described with only small modifications of standard x-ray tube geometry.

A spatio-temporal detective quantum efficiency and its application to fluoroscopic systems

PURPOSE

Fluoroscopic x-ray imaging systems are used extensively in spatio-temporal detection tasks and require a spatio-temporal description of system performance. No accepted metric exists that describes spatio-temporal fluoroscopic performance. The detective quantum efficiency (DQE) is a metric widely used in radiography to quantify system performance and as a surrogate measure of patient "dose efficiency". It has been applied previously to fluoroscopic systems with the introduction of a temporal correction factor. However, the use of a temporally-corrected DQE does not provide system temporal information and it is only valid under specific conditions, many of which are not likely to be satisfied by suboptimal systems. The authors propose a spatio-temporal DQE that describes performance in both space and time and is applicable to all spatio-temporal quantum-based imaging systems.

METHODS

The authors define a spatio-temporal DQE (two spatial-frequency axes and one temporal-frequency axis) in terms of a small-signal spatio-temporal modulation transfer function (MTF) and spatio-temporal noise power spectrum (NPS). Measurements were made on an x-ray image intensifier-based bench-top system using continuous fluoroscopy with an RQA-5 beam at 3.9 microR/frame and hardened 50 kVp beam (0.8 mm Cu filtration added) at 1.9 microR/frame.

RESULTS

A zero-frequency DQE value of 0.64 was measured under both conditions. Nonideal performance was noted at both larger spatial and temporal frequencies; DQE values decreased by approximately 50% at the cutoff temporal frequency of 15 Hz.

CONCLUSIONS

The spatio-temporal DQE enables measurements of decreased temporal system performance at larger temporal frequencies analogous to previous measurements of decreased (spatial) performance. This marks the first time that system performance and dose efficiency in both space and time have been measured on a fluoroscopic system using DQE and is the first step toward the generalized use of DQE on clinical fluoroscopic systems.

[An investigation of the subjective assessment of fluoroscopic imaging using instantaneous detectability]

Instantaneous detectability was introduced in a method for the subjective assessment of fluoroscopic images, particularly in the interventional radiology (IVR) procedure. Quantitatively, instantaneous detectability was obtained by measuring the time required for observers to detect the tip of a linear pattern, such as a guide wire, just after the image is displayed. Dynamic images used in this measurement, which mimic degraded fluoroscopic images, were created in the computer simulation software by adding a low-contrast linear pattern to a noisy background image. Radiological technologists and students in the faculty of computer engineering participated in the assessment, and all measurements were performed using a personal computer system. Even if the contrast-to-noise ratio was identical, instantaneous detectability was remarkably increased when the background noise was dominated by higher frequency components. Also, the sign test suggested that a frame rate of 30 f/s significantly improved detectability compared to a frame rate of 15 f/s. These results enable us to discuss new possibilities for image processing and the optimization of system performance. Although the standard deviation of the measured inter- and intra-observer data was large, statistical significance should be suitably examined by a paired-comparison like the sign test, which will be one of the important analyzers in experiments investigating human performance.

A small-signal approach to temporal modulation transfer functions with exposure-rate dependence and its application to fluoroscopic detective quantum efficiency

The detective quantum efficiency (DQE) is a metric widely used in radiography to quantify system performance and as a surrogate measure of patient "dose efficiency." It has been applied previously to fluoroscopic systems with the introduction of a temporal correction factor. Calculation of this correction factor relies on measurements of the temporal modulation transfer function (MTF). However, the temporal MTF is often exposure-rate dependent, violating a necessary Fourier linearity requirement. The authors show that a Fourier analysis is appropriate for fluoroscopic systems if a "small-signal" approach is used. Using a semitransparent edge, a lag-corrected DQE is described and measured for an x-ray image intensifier-based fluoroscopic system under continuous (non-pulsed) exposure conditions. It was found that results were equivalent for both rising and falling-edge profiles independent of edge attenuation when effective attenuation was in the range of 0.1-0.6. This suggests that this range is appropriate for measuring the small-signal temporal MTF. In general, lag was greatest at low exposure rates. It was also found that results obtained using a falling-edge profile with a radiopaque edge were equivalent to the small-signal results for the test system. If this result is found to be true generally, it removes the need for the small-signal approach. Lag-corrected DQE values were validated by comparison with radiographic DQE values obtained using very long exposures under the same conditions. Lag was observed to inflate DQE measurements by up to 50% when ignored.

Noise reduction in low-dose x-ray fluoroscopy for image-guided radiation therapy

PURPOSE

To improve the quality of low-dose X-ray fluoroscopic images using statistics-based restoration algorithm so that the patient fluoroscopy can be performed with reduced radiation dose.

METHOD AND MATERIALS

Noise in the low-dose fluoroscopy was suppressed by temporal and spatial filtering. The temporal correlation among neighboring frames was considered by the Karhunen-Loève (KL) transform (i.e., principal component analysis). After the KL transform, the selected neighboring frames of fluoroscopy were decomposed to uncorrelated and ordered principal components. For each KL component, a penalized weighted least-squares (PWLS) objective function was constructed to restore the ideal image. The penalty was chosen as anisotropic quadratic, and the penalty parameter in each KL component was inversely proportional to its corresponding eigenvalue. Smaller KL eigenvalue is associated with the KL component of lower signal-to-noise ratio (SNR), and a larger penalty parameter should be used for such KL component. The low-dose fluoroscopic images were acquired using a Varian Acuity simulator. A quality assurance phantom and an anthropomorphic chest phantom were used to evaluate the presented algorithm.

RESULTS

In the images restored by the proposed KL domain PWLS algorithm, noise is greatly suppressed, whereas fine structures are well preserved. Average improvement rate of SNR is 75% among selected regions of interest. Comparison studies with traditional techniques, such as the mean and median filters, show that the proposed algorithm is advantageous in terms of structure preservation.

CONCLUSIONS

The proposed noise reduction algorithm can significantly improve the quality of low-dose X-ray fluoroscopic image and allows for dose reduction in X-ray fluoroscopy.

A moving slanted-edge method to measure the temporal modulation transfer function of fluoroscopic systems

Lag in fluoroscopic systems introduces a frame-averaging effect that reduces measurements of image noise and incorrectly inflates measurements of the detective quantum efficiency (DQE). A correction can be implemented based on measurements of the temporal modulation transfer function (MTF). We introduce a method of measuring the temporal MTF under fluoroscopic conditions using a moving slanted edge, a generalization of the slanted-edge method used to measure the (spatial) MTF, providing the temporal MTF of the entire imaging system. The method uses a single x-ray exposure, constant edge velocity, and assumes spatial and temporal blurring are separable. The method was validated on a laboratory x-ray image intensifier (XRII) system by comparison with direct measurements of the XRII optical response, showing excellent agreement over the entire frequency range tested (+/- 100 Hz). With proper access to linearized data and continuous fluoroscopy, this method can be implemented in a clinical setting on both XRII and flat-panel detectors. It is shown that the temporal MTF of the CsI-based validation system is a function of exposure rate. The rising-edge response showed more lag than the falling edge, and the temporal MTF decreased with decreasing exposure rate. It is suggested that a small-signal approach, in which the range of exposure rates is restricted to a linear range by using a semitransparent moving edge, would be appropriate for measuring the DQE of these systems.

An algorithm for tracking microcatheters in fluoroscopy

Currently, a large number of endovascular interventions are performed for treatment of intracranial aneurysms. For these treatments, correct positioning of microcatheter tips, microguide wire tips, or coils is essential. Techniques to detect such devices may facilitate endovascular interventions. In this paper, we describe an algorithm for tracking of microcatheter tips during fluoroscopically guided neuroendovascular interventions. A sequence of fluoroscopic images (1,024 x 1,024 x 12 bits) was acquired using a C-arm angiography system as a microcatheter was passed through a carotid phantom which was on top of a head phantom. The carotid phantom was a silicone cylinder containing a simulated vessel with the shape and curvatures of the internal carotid artery. The head phantom consisted of a human skull and tissue-equivalent material. To detect the microcatheter in a given fluoroscopic frame, a background image consisting of an average of the four previous frames is subtracted from the current frame, the resulting image is filtered using a matched filter, and the position of maximum intensity in the filtered image is taken as the catheter tip position in the current frame. The distance between the tracked position and the correct position (error distance) was measured in each of the fluoroscopic images. The mean and standard deviation of the error distance values were 0.277 mm (1.59 pixels) and 0.26 mm (1.5 pixels), respectively. The error distance was less than 3 pixels in the 93.0% frames. Although the algorithm intermittently failed to correctly detect the catheter, the algorithm recovered the catheter in subsequent frames.

An observer study methodology for evaluating detection of motion abnormalities in gated myocardial perfusion SPECT

To address the task of detecting nonischemic motion abnormalities from animated displays of gated myocardial perfusion single photon emission computed tomography data, we performed an observer study to evaluate the difference in detection performance between gating to 8 and 16 frames. Images were created from the NCAT mathematical phantom with a realistic heart simulating hypokinetic motion in the left lateral wall. Realistic noise-free projection data were simulated for both normal and defective hearts to obtain 16 frames for the cardiac cycle. Poisson noise was then simulated for each frame to create 50 realizations of each heart, All datasets were processed in two ways: reconstructed as a 16-frame set, and collapsed to 8 frames and reconstructed. Ten observers viewed the cardiac images animated with a realistic real-time frame rate. Observers trained on 100 images and tested on 100 images, rating their confidence on the presence of a motion defect on a continuous scale. None of the observers showed a significant difference in performance between the two gating methods. The 95% confidence interval on the difference in areas under the ROC curve (Az8 - Az16) was -0.029-0.085. Our test did not find a significant difference in detection performance between 8-frame gating and 16-frame gating. We conclude that, for the task of detecting abnormal motion, increasing the number of gated frames from 8 to 16 offers no apparent advantage.

Quantification of motion unsharpness in digital fluoroscopy

The objectives of this work were first to develop a convenient method to quantify persistence in digital fluoroscopy systems, then to quantify the effect of variable temporal averaging on the detection of moving low-contrast test details within digital fluoroscopic and pulsed fluoroscopic images. The results were analysed to clarify the relationship between the optimum persistence required to see the lowest contrast for circular test details for a range of diameters and their speed of movement. The optimum persistence values obtained are compared with the limited data available on speeds of movement of patient organs during fluoroscopy. It is tentatively concluded that for imaging the abdomen, the optimum imaging system persistence time constant is approximately 0.15 s. For the much greater speeds associated with cardiac motion, no additional frame averaging is necessary, i.e. just the persistence provided by the observer's visual system appears to be optimal for small objects.

Quantitative image quality evaluation of pixel-binning in a flat-panel detector for x-ray fluoroscopy

X-ray fluoroscopy places stringent design requirements on new flat-panel (FP) detectors, requiring both low-noise electronics and high data transfer rates. Pixel-binning, wherein data from more that one detector pixel are collected simultaneously, not only lowers the data transfer rate but also increases x-ray counts and pixel signal-to-noise ratio (SNR). In this study, we quantitatively assessed image quality of image sequences from four acquisition methods; no-binning and three types of binning; in synthetic images using a clinically relevant task of detecting an extended guidewire in a four-alternative forced-choice paradigm. Binning methods were conventional data-line (D) and gate-line (G) binning, and a novel method in which alternate frames in an image sequence used D and G binning. Two detector orientations placed the data lines either parallel or perpendicular to the guide wire. At a low exposure of 0.6 microR (1.548 x 10(-10) C/kg) per frame, irrespective of detector orientation, D binning with its reduced electronic noise was significantly (p<0.1) better than the other acquisition methods. On average, alternate binning performed better than G binning. At a higher exposure of 4.0 microR (10.32 x 10(-10) C/kg) per frame, with data lines parallel to the guidewire, detection with D binning was significantly (p<0.1) better than G binning. However, with data lines perpendicular to the guidewire, G binning was significantly (p<0.1) better than D binning because the partial area effect was reduced. Alternate binning was the best binning method when results were averaged over both orientations, and it was as good as the best binning method at either orientation. In addition, at low and high exposures, alternate binning gave a temporally fused image with a smooth guidewire, an important image quality feature not assessed in a detection experiment. While at high exposure, detection with no binning was as good, or better, than the best binning method, it might be impractical at fluoroscopy imaging rates. A computational observer model based on signal detection theory successfully fit data and was used to predict effects of similar acquisition methods. Results from this study suggest the use of exposure-dependent detector binning in fluoroscopy that switches between D binning and alternate binning at low and high exposures, respectively.

Optimization of variable temporal averaging in digital fluoroscopy

In modern X-ray fluoroscopy systems, the amount of temporal averaging (i.e. persistence) applied to the image is often user selectable. The objective of this work is to quantify the effect of variable temporal averaging on the detection of low contrast test objects moving at a range of known speeds within the digital fluoroscopic image. An image intensifier system with a short-persistence television camera was used to record image sequences of a moving threshold contrast-detail diameter test object onto broadcast-standard U-matic videotape. The image sequences were replayed through an image processing system allowing different amounts of temporal averaging to be applied. The test images were scored by an experienced observer. The temporal averaging time constants produced by the added image processing were measured using a method based on noise correlation. Results are presented showing the trends of threshold contrast with test detail diameter and movement speed. The optimum value of temporal averaging time constant is presented as a function of detail diameter for a range of speeds. By comparison with the limited information available in the literature on organ movement, it is tentatively concluded that for the organ movement speeds expected in the abdomen, the optimum imaging system persistence time constant should be approximately 0.15 s. For the much greater speeds associated with cardiac motion no additional frame averaging, i.e. just the persistence provided by the observer's visual system, appears to be optimal.

Quantitative assessment of image quality enhancement due to unsharp-mask processing in x-ray fluoroscopy

Spatial unsharp-mask processing and its variants are commonly used in x-ray radiography to enhance image contrast. We investigated the effect of three unsharp-masking filter kernels of different sizes on the detection of an advanced guidewire tip in simulated x-ray fluoroscopy image sequences. To isolate the effect of visual temporal processing, we repeated the experiments on single images. Filter gains were selected so that all three kernels increased the contrast of a 0.018-in. (0.457-mm) guidewire by a factor of 2 but had different effects on image noise and signal profiles. There was no statistically significant effect of unsharp masking on human-observer performance in single images. However, all three kernels significantly improved average performance in image sequences, and the guidewire contrast required for detection was reduced by 32%-40%. A prewhitening channelized observer model predicted the disparity between sequences and single images and fitted measurements at different kernel sizes well. A nonprewhitening observer model did not. We conclude that unsharp masking is a simple and effective method of improving guidewire visualization in fluoroscopically guided interventional procedures and that quantitative image quality studies are essential for evaluation of image-processing techniques in sequences such as x-ray fluoroscopy.

Image quality evaluation of flat panel and image intensifier digital magnification in x-ray fluoroscopy

Interventional devices used in radiology often have dimensions on the order of a pixel, and radiologists resort to image magnification to better visualize such small devices. Traditional image intensifier (II) systems use analog magnification with x-ray exposure inversely proportional to the area of field of view (FOV) so as to maintain light output for the camera. Analog magnification is impossible with flat panel (FP) detectors, and images must be magnified using digital interpolation that does not reduce the pixel partial area effect for small devices. We quantitatively investigated image quality of digital and analog magnification using a clinically relevant task of detecting a partially deployed stent in x-ray fluoroscopy image sequences that were created using realistic detector models. Using the standard exposure strategy for II analog magnification, exposure was increased from a nominal 43.65 nGy (5.0 microR) per frame at 23 cm FOV to 79.9 nGy (9.15 microR) per frame and 117.81 nGy (13.49 microR) per frame at 17 cm and 14 cm FOV, respectively. Contrast sensitivity improved significantly (p<0.1) by 43.5+/-6.5% and 64.1+/-7.3% with the 17 cm and 14 cm FOV, respectively. Exposure for digitally magnified images was varied in an adaptive forced choice experiment so as to match performance with II analog magnification. For digital magnification, bilinear interpolation was used to give magnified stents sizes equivalent to those in the analog magnified images. For equivalent image quality, FP required 34.87+/-2.59, 80.16+/-5.37, and 84.08+/-5.59 nGy per frame at normal, and the two magnification modes, respectively. Hence, FP with digital magnification gives significant (p<0.1) dose savings of 20+/-6% and 27+/-5% at the normal and highest magnification modes, respectively. Digitally magnified II images required exposures of 110.85+/-8.07 and 103.34+/-5.90 nGy per frame for the two magnifications levels, respectively, giving no significant (p>0.1) dose savings. A spatiotemporal human observer model based on signal detection theory successfully predicted the human data and was used to predict other conditions associated with image magnification. Model predictions quantitatively showed that magnification is most useful when signal size is relatively small and that FP digital magnification can improve image quality for the stent deployment task without increasing exposure. In conclusion, the results show that FP digital magnification can be useful and dose efficient as compared to analog magnification.

Hyperefficient detection of targets in noisy images

We compared human detection of visual targets in noisy images with that of a theoretically optimum matched filter. Using a small thin target with vertically aligned markers, we obtained hyperefficient detection as high as 91% as compared with the theoretical optimum, a value far exceeding the 30-50% value typically reported. When the markers were removed, detection efficiencies degraded to an average of 27%, even though subjects were aware that the target was always placed in the center of a reasonably small panel. Using a nine-alternative forced-choice experiment, we compared detection by human observers with a matched-filter computational observer on a trial-by-trial basis. With the markers present, when humans missed the correct panel, they most often chose the panel with the second-highest decision variable output from the computational observer, suggesting that the template-matching model is a good one. To model results without the markers, we included location uncertainty and additional noise sources in the template matching of the computational observer. A location uncertainty of only 1 pixel, corresponding to a retinal distance of approximately 12 microm, a dimension of the order of the size of the receptive field of photoreceptors, explained the psychometric data. With the marker present, the model suggests that hyperefficient detection is obtained by limiting target location uncertainty to <6 microm. Together these results give important new insights into human visual detection mechanisms.

Quantitative image quality analysis of a nonlinear spatio-temporal filter

Digital temporal and spatial filtering of fluoroscopic image sequences can be used to improve the quality of images acquired at low X-ray exposure. In this study, we characterized a nonlinear edge preserving, spatio-temporal noise reduction filter, the bidirectional multistage (BMS) median filter of Arce (1991). To assess image quality, signal detection and discrimination experiments were performed on stationary targets using a four-alternative forced-choice paradigm. A measure of detectability, d', was obtained for filtered and unfiltered noisy image sequences at different signal amplitudes. Filtering gave statistically significant, average d' improvements of 20% (detection) and 31% (discrimination). A nonprewhitening detection model modified to include the human spatio-temporal visual system contrast-sensitivity underestimated enhancement, predicting an improvement of 6%. Pixel noise standard deviation, a commonly applied image quality measure, greatly overestimated effectiveness giving 67% improvement in d'. We conclude that human testing is required to evaluate the filter effectiveness and that human perception models must be improved to account for the spatio-temporal filtering of image sequences.

Effect of inherent location uncertainty on detection of stationary targets in noisy image sequences

The effect of inherent location uncertainty on the detection of stationary targets was determined in noisy image sequences. Targets were thick and thin projected cylinders mimicking arteries, catheters, and guide wires in medical imaging x-ray fluoroscopy. With the use of an adaptive forced-choice method, detection contrast sensitivity (the inverse of contrast) was measured both with and without marker cues that directed the attention of observers to the target location. With the probability correct clamped at 80%, contrast sensitivity increased an average of 77% when the marker was added to the thin-cylinder target. There was an insignificant effect on the thick cylinder. The large enhancement with the thin cylinder was obtained even though the target was located exactly in the center of a small panel, giving observers the impression that it was well localized. Psychometric functions consisting of d' plotted as a function of the square root of the signal-energy-to-noise-ratio gave a positive x intercept for the case of the thin cylinder without a marker. This x intercept, characteristic of uncertainty in other types of detection experiments, disappeared when the marker was added or when the thick cylinder was used. Inherent location uncertainty was further characterized by using four different markers with varying proximity to the target. Visual detection by human observers increased monotonically as the markers better localized the target. Human performance was modeled as a matched-filter detector with an uncertainty in the placement of the template. The removal of a location cue was modeled by introducing a location uncertainty of approximately equals 0.4 mm on the display device or only 7 microm on the retina, a size on the order of a single photoreceptor field. We conclude that detection is affected by target location uncertainty on the order of cellular dimensions, an observation with important implications for detection mechanisms in humans. In medical imaging, the results argue strongly for inclusion of high-contrast visualization markers on catheters and other interventional devices.

Dose efficiency and low-contrast detectability of an amorphous silicon x-ray detector for digital radiography

The effect of dose reduction on low-contrast detectability is investigated theoretically and experimentally for a production grade amorphous silicon (a-Si) x-ray detector and compared with a standard thoracic screen-film combination. A non-prewhitening matched filter observer model modified to include a spatial response function and internal noise for the human visual system (HVS) is used to calculate a signal-to-noise ratio (SNR) related to object detectability. Other inputs to the SNR calculation are the detective quantum efficiency (DQE) and the modulation transfer function (MTF) of the imaging system. Besides threshold detectability, the model predicts the equivalent perception dose ratio (EPDR), which is the fraction of the screen film exposure for which the digital detector provides equal detectability. Images of a contrast-detail phantom are obtained with the digital detector at dose levels corresponding to 27%, 41%, 63% and 100% of the dose used for screen-film. The images are used in a four-alternative forced choice (4-AFC) observer perception study in order to measure threshold detectability. A statistically significant improvement in contrast detectability is measured with the digital detector at 100% and 63% of the screen-film dose. There is no statistical difference between screen-film and digital at 41% of the dose. On average, the experimental EPDR is 44%, which agrees well with the model prediction of 40%.