British Journal of Radiology (2003) 76, 648-652
© 2003 British Institute of Radiology
doi: 10.1259/bjr/25693100
A practical approach to soft-copy display consistency for PC-based review workstations
S E Jervis, MSc
and
D S Brettle, MSc
Department of Medical Physics & Engineering, Leeds Teaching Hospitals NHS Trust, St James's University Hospital, Leeds LS9 7TF, UK
 |
Abstract
|
|---|
The use of non-permanent, digital image display, i.e. soft-copy display, is increasing within hospitals due to the growth in the use of digital modalities and picture archiving and communication systems (PACS). Non-dedicated image review using standard PCs is being employed as a cost-effective method of image access. These workstations do not have specialized display systems and are likely to suffer from inconsistent image presentation. The Digital Imaging and Communications in Medicine (DICOM) Working Group 11 has developed a display function standard (part 14) to standardize the display of grey scale images. Although this standard is starting to be adopted by manufacturers of proprietary reporting systems it is not easily applied to the existing number of non-dedicated, PC-based review systems. The aim of this work was to investigate whether display consistency could be achieved simply and reproducibly on these systems, outside of the DICOM standard: part 14, by adjusting monitor brightness and contrast settings and using the Society of Motion Picture and Television Engineers (SMPTE) digital test pattern. The study showed that by adjusting the brightness and contrast settings alone it was possible to approximate the display characteristic curves to the grey scale standard display function (GSDF) defined in the DICOM standard: part 14, but only at unacceptably low luminances. Intradisplay and interdisplay consistency could be achieved using a simple monitor set-up procedure and the SMPTE test pattern.
|
Introduction
|
|---|
The increased use of digital modalities and picture archiving and communication systems (PACS) in the last decade has resulted in a rapid growth in the number of reporting and clinical review workstations used for the non-permanent, digital image display, i.e. soft copy display, of radiological images. This growth brings with it an increased potential for display variations that could result in non-optimized presentation of radiological images. Quality assurance (QA) systems can be implemented to ensure consistent display, particularly when systems are installed within QA active environments, such as radiology departments. Image display systems outside the direct control of the radiology department, such as teleradiology systems and systems that use web-based review software, present more of a problem. These systems tend to be non-dedicated, standard PC-based review systems, which are less expensive and have lower display specifications than dedicated reporting workstations.
Display consistency can be separated into two components: intradisplay and interdisplay consistency. The former is affected by the set-up of the display, e.g. contrast, brightness, display frequency, grey scale bit-depth, resolution, and the performance of the monitor, e.g. phosphor degradation, cathode wear-out, cut-off voltage drift. Intradisplay consistency can be maximized by optimal base-line set-up and an appropriate QA programme. Interdisplay consistency is affected by display set-up and monitor performance, and also by the inherent grey scale response (the characteristic curve) of the display. The inherent grey scale response can vary within monitors of the same make and model. Interdisplay consistency can be controlled by mapping the display characteristic curve through a look-up table (LUT) generated using a calibrated standardized display function. To maintain a standardized output using such a method the display settings, e.g. contrast and brightness, must remain constant and as such there is still a requirement for routine QA.
The Royal College of Radiologists (RCR) and the American College of Radiology (ACR) guidelines for teleradiology and digital image data management [1–3] advocate the implementation of a QA procedure for displays. There is supporting work to guide the development of such a programme [4–7]. However, the potentially large numbers of displays makes routine QA resource intensive and consequently it is often a low priority area, despite the fact that poor performance of the display system can undermine the entire imaging chain.
To address the issue of interdisplay consistency, the Digital Imaging and Communications in Medicine (DICOM) Working Group 11 has developed a display function standard (part 14) [8]. This standard aims to define mathematically a grey scale standard display function (GSDF) for all image presentation systems which will provide similarity in grey scale perception for a given image between display systems of different luminance. It also aims to facilitate good use of the digital driving levels (DDL) of those systems, where a DDL is the digital input to a display system which produces a given luminance value. The GSDF uses the concept of perceptual linearization to yield a grey scale in which the display driving levels produce perceptually-equivalent luminance steps over the total luminance range for a specific test pattern. For a given display system, conformance with the GSDF can be achieved by measuring the characteristic curve of the system, using a test pattern detailed in the standard, and generating a LUT to map the measured characteristic curve to the GSDF. Some manufacturers are starting to incorporate the GSDF into the configuration and the calibration of their systems. However, for the majority of non-dedicated, standard PC-based review systems the implementation of the GSDF will be a difficult, and in some cases impossible, task.
The aim of this study was to investigate whether the display output of PC-based systems could be standardized, without implementing software LUT calibration to the DICOM standard: part 14 GSDF. Initial work was carried out to investigate the effect of the brightness and contrast settings on the display characteristic curves and whether the alteration of these settings alone could be used to approximate the curves to the GSDF. The second phase of the work was carried out to assess whether intradisplay and interdisplay consistency could be achieved using a digital test image and a simple set-up protocol. The test image used was the Society of Motion Picture and Television Engineers (SMPTE) digital test pattern [9]. This is a test pattern that can be used to provide measures of system performance, such as spatial resolution and contrast sensitivity. Various authors have indicated the usefulness of this test pattern for QA monitoring [4–7, 10].
|
Methods and materials
|
|---|
Table 1
contains the display specifications of the four colour PC displays used in the work.
The effect of varying the contrast and brightness on the display characteristic curves was investigated using the two Iiyama (Stevenage, UK) monitors (displays A and C), as the contrast and brightness values were digitally displayed on the set-up menu. The characteristic curves of the displays were measured using a test pattern consisting of a square measurement field (10% of the total number of pixels displayed by the system) displayed centrally on a full-screen uniform background of 20% of maximum luminance and a photometer (Unfors Light-O-Meter, Unfors Instruments, Billdal, Sweden) as described in the DICOM standard: part 14 [8]. The characteristic curves were measured for all combinations of 100%, 75% and 50% contrast and brightness levels.
All four displays were used to investigate the reproducibility of monitor set-up using the SMPTE test pattern. The ambient lighting was minimized and the resulting illuminance measured at the monitor face using a photometer. The SMPTE test pattern was displayed full-screen with the contrast level set to approximately 50%. The brightness was adjusted, using the 0/5% small contrast changes square, until the outside 0% region of the square just changed from grey to black, without losing contrast with the centre 5% region. The contrast was altered using the 95/100% small contrast changes square to give a relatively bright white level without over-saturating and losing the contrast between the outer 100% and inner 95% regions. The characteristic curves of the displays were measured using the DICOM standard: part 14 test pattern and a photometer. The characteristic curves were measured for each display a total of three times, resetting the contrast and brightness controls on each occasion and following the set-up described above. In the cases of displays A and B the repeat measurements were consecutive. The measurements carried out on displays C and D took place on two separate occasions with a time interval of 1 week between them.
To investigate variations in the shapes of the characteristic curves for the four displays the curve measurements were repeated with the minimum and maximum luminance levels fixed at 0.1 cdm-2 and 90 cdm-2, respectively. These values were found to be typical of the black and peak white luminances of the four displays tested, after optimization with the SMPTE test pattern.
|
Results
|
|---|
Figure 1 shows the variation in the characteristic curve of display A with contrast and brightness. Plotted on each is the GSDF. The GSDF is defined over a large range of luminance values (0.05–3916 cdm-2), and in each case the section of the GSDF between the maximum and minimum measured luminances is shown. Similar curves were measured for display C.
View larger version (27K):
[in this window]
[in a new window]
|
Figure 1. Variation in the characteristic curve (luminance against normalized digital driving level (DDL)) of display A with contrast and brightness (crosses), with corresponding sections of the grey scale standard display function (GSDF) curve (solid curve).
|
|
Figure 2 shows the measurements of the characteristic curves for each display when repeatably set-up using the SMPTE test pattern with the set-up protocol. Plotted on each is the section of the GSDF constrained to the average maximum and minimum measured luminance values.
View larger version (27K):
[in this window]
[in a new window]
|
Figure 2. Repeat measurements of the characteristic curves for each display [(a) display A; (b) display B; (c) display C; (d) display D], plotted with corresponding sections of the grey scale standard display function (GSDF) curve constrained to the average maximum and minimum measured luminances. DDL, digital driving level.
|
|
Constraining the minimum and maximum luminance values to 0.1 cdm-2 and 90 cdm-2, respectively, for each of the displays resulted in the characteristic curves shown in Figure 3.
View larger version (15K):
[in this window]
[in a new window]
|
Figure 3. Characteristic curves measured for all the displays with the maximum and minimum luminance values constrained to the same values, shown with the corresponding sections of the grey scale standard display function (GSDF) curve. DDL, digital driving level.
|
|
Figure 4 shows the relationship between ambient lighting conditions (average illuminance) and the average minimum and maximum luminance values measured for each display, after set-up using the SMPTE test pattern and set-up protocol. Linear regression performed on the data revealed statistically significant linear relationships between average illuminance and average minimum luminance with a slope of 0.007 (95% CI: 0.003 to 0.010; p-value<0.02) and average illuminance and average maximum luminance with a slope of 0.7 (95% CI: 0.2 to 1.3; p-value<0.03).
View larger version (10K):
[in this window]
[in a new window]
|
Figure 4. Average illuminance plotted against (a) average minimum luminance and (b) average maximum luminance values for each display.
|
|
|
Discussion
|
|---|
The display characteristic curves were found to vary in a predictable manner with brightness and contrast, with the brightness control having the largest effect on the minimum luminance value and the contrast control on the maximum luminance value. Figure 1
shows that as contrast and brightness decreased the characteristic curve tended towards the GSDF, and hence a perceptually linearized curve. This was only possible at unacceptably low luminances. For example in the case of display A the best fit was attained at a maximum luminance of 18 cdm-2. The ACR has published two standards [2, 3] specifying a minimum luminance of 171 cdm-2 for monitors used for reporting. For image review purposes no such display criteria exist and it is acceptable to use non-dedicated, PC-based displays for this purpose.
The intradisplay consistency of the displays, when set up using the SMPTE test pattern and standard protocol, depended on the reproducibility in setting the maximum and minimum luminance values. The resulting intradisplay characteristic curves (Figure 2) were well matched, indicating a high level of reproducibility in setting-up the monitors. The curvatures of the measured characteristic curves were similar and were less than that of the GSDF. The gradient of the GSDF increases with luminance and therefore the displays set with the lower maximum luminance values, e.g. display A, were found to have a better fit to the GSDF.
The variation in the average characteristic curves measured for each display indicated that interdisplay consistency was poorer than intradisplay consistency. The interdisplay consistency depended on the intrinsic shape of the display characteristic curves and the set-up of the maximum and minimum luminance values. Fixing the minimum and maximum luminance values of the displays showed that the intrinsic shapes of the characteristic curves were well-matched, as is shown in Figure 3. Some of the set-up imprecision can be accounted for by the effect of the ambient lighting conditions on the minimum and maximum luminance values of the characteristic curves. Each display was housed in a separate room with varying ambient lighting conditions, and correlations were found between the average illuminance and the resulting average minimum luminance (black) and average maximum luminance (peak white) values set up on the displays, as shown in Figure 4. The relationship between the illuminance level and the set-up of the characteristic curve using the SMPTE test pattern highlights the importance of standardizing viewing conditions and setting-up the display with the ambient lighting conditions under which it will be used.
Overall the results show that it is possible to achieve a level of display consistency on PC-based review workstations simply by using the SMPTE test pattern and a standardized set-up protocol. Although the accuracy of this method will not match that produced by the implementation of the GSDF curve it will result in the display of perceptually similar images. This work comprises only one element of an adequate soft-copy QA programme. Other areas that such a programme should monitor include temporal luminance stability, spatial resolution and screen geometry. Other authors [4–7] have given recommendations for practical QA procedures and how often they should be carried out. Soft-copy QA will become more important as display/communications technology converges, resulting in the merging of the functions of review and reporting workstations.
|
Conclusion
|
|---|
With the growth of distributed, non-dedicated, PC-based review workstations it is becoming increasingly important to control display settings. This work has shown that in systems in which software LUT calibration to the GSDF is not implemented, it is not possible to standardize displays by approximating display characteristic curves to the GSDF using the brightness and contrast settings, whilst maintaining acceptable levels of luminance. However, display consistency can be achieved outside of the DICOM standard: part 14 using a simple set-up protocol and the SMPTE test pattern.
Received for publication August 22, 2002.
Revision received April 18, 2003.
Accepted for publication May 29, 2003.
|
References
|
|---|
- Clinical Radiology and Electronic Records, The Royal College of Radiologists 2002.
- ACR Standard for Digital Image Data Management, ACR Standards 1998 (rev. 2001).
- ACR Standard for Teleradiology, ACR Standards 1994 (rev. 1998).
- Groth DS, Bernatz SN, Fetterly KA, Hangiandreou NJ. Cathode ray tube quality control and acceptance testing program: initial results for clinical PACS displays. Radiographics 2001;21:719–32.[Abstract/Free Full Text]
- Roehrig H, Willis CE, Damento MA. Characterization of monochrome CRT display systems in the field. J Digit Imaging 1999;12:152–65.
- Parsons DM, Kim Y, Haynor DR. Quality control of cathode-ray tube monitors for medical imaging using a simple photometer. J Digit Imaging 1995;8:10–20.[Medline]
- Gray JE, et al. Test pattern for video displays and hard-copy cameras. Radiology 1985;154:519–27.[Abstract/Free Full Text]
- Digital Imaging and Communications in Medicine (DICOM) Part 14: Grayscale Standard Display Function. National Electrical Manufacturers Association (NEMA) PS 3.14-2001.
- SMPTE Recommended Practice: Specification for Medical Diagnostic Imaging Test Pattern for Television Monitors and Hard-Copy Recording Cameras. Society of Motion Picture and Television Engineers, 1991.
- Nawfel RD, Chan KH, Wagenaar DJ, Judy PF. Evaluation of video gray-scale display. Med Phys 1992;19:561–7.[CrossRef][Medline]
This article has been cited by other articles:
|
|
|
|
 
D S Brettle
Display considerations for hospital-wide viewing of soft copy images
Br. J. Radiol.,
July 1, 2007;
80(955):
503 - 507.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
K. J. Robson and C. J. Kotre
Pilot study into optimisation of viewing conditions for electronically displayed images
Radiat Prot Dosimetry,
December 1, 2005;
117(1-3):
298 - 303.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
D S Brettle and S E Bacon
A method for verified access when using soft copy display
Br. J. Radiol.,
August 1, 2005;
78(932):
749 - 751.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
W. Berkhout, D. Beuger, G. Sanderink, and P. van der Stelt
The dynamic range of digital radiographic systems: dose reduction or risk of overexposure?
Dentomaxillofac. Radiol.,
January 1, 2004;
33(1):
1 - 5.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
