British Journal of Radiology (2005) 78, 749-751
© 2005 British Institute of Radiology
doi: 10.1259/bjr/19733434
A method for verified access when using soft copy display
D S Brettle, PhD
and
S E Bacon, MSc
Department of Medical Physics & Engineering, Leeds Teaching Hospitals NHS Trust, Leeds General Infirmary, Leeds LS1 3EX, UK
 |
Abstract
|
|---|
Soft copy display is a rapidly developing area. To date, most soft copy systems can be classed by their application, e.g. review or reporting. With technology convergence this distinction is becoming less defined by the hardware and more defined by the software functionality. Although it is accepted that routine quality assurance should be conducted on soft copy monitors, this would be logistically difficult to achieve if any monitor within a hospital could be used for image review or reporting. This work proposes a simple psychophysical check to ensure optimal display performance before viewing software can be run. This is in the form of a challenge/response code constructed from letters just above the threshold of detection. This verified login would act as a portal to launching the image viewing software. The developed system was tested on three different types of monitor and five observers. Results indicate that the verified login was able to control access for displays below the optimal settings but was not as sensitive for adjustments above the optimum. However it is believed this is still of value as the lower presentation will compress the display gamma curve and reduce detail contrast. It also provides a minimum level of audit and quality control that might otherwise be missing.
|
Introduction
|
|---|
The use of soft copy display is increasing in UK hospitals due to the expansion of digital imaging systems and picture archiving and communications systems (PACS). Images can now be viewed on a range of display systems, from dedicated multi-monitor workstations to desktop PCs and home computers using web browser software. Although the display is the last part of a costly imaging chain, it is often overlooked in routine quality assurance (QA) programmes, even though there is evidence that such a system is required [1–3]. Some high end monitors do have integrated QA systems, but at present this level of control is limited to reporting workstations due to the cost involved. Consistent display is difficult to control outside of a routine QA environment as the users often have ready access to fundamental display settings, e.g. contrast and brightness, which can dramatically affect grey scale rendition. Alterations to the brightness settings demonstrate a larger effect at the low end of the digital signal, whereas the high end is effected more significantly by alterations in the contrast [4].
Although it is accepted that routine QA is required to maintain optimal display characteristics, providing this service to large numbers of disparate systems would be cost prohibitive. With modern PACS systems, web-based image access is an integral component. This means that potentially every PC in a hospital can be used for viewing medical images. Even if it were logistically feasible to perform QA on each monitor there is the additional problem of preventing the display settings from being changed between QA tests. For example, a user may adjust the image contrast and brightness using the monitor controls instead of using the software tools. The next user would be unaware of this and consequently would view their images on a sub-optimal display. Currently the only solution to this problem is to prohibit user access to monitor settings. In the ideal world, soft copy display QA should be implemented locally before each viewing session.
It is proposed that a mechanism for verifying display brightness and contrast presentation, implemented every time the display software is accessed, would provide a minimum failsafe against sub-optimal display settings when viewing medical images.
|
Method
|
|---|
The proposed method for restricting access to sub-optimal display systems is through a verified login interface program. Initially this interface is run instead of the viewing software and issues a challenge code constructed from random letters at pre-determined threshold grey level values. Each letter of the code is superimposed on a different background intensity of 0% (black), 50% (mid 1: negative contrast letter), 50% (mid 2: positive contrast letter) and 100% (white) of maximum grey level. The interface is displayed on top of a full field, uniform background window set at 20% of maximum luminance to replicate the conditions used for the DICOM part 14 test pattern [5]. An example of the verified login display is shown in Figure 1
. If the user correctly repeats the code the viewing software is run. If not, the user is prompted to set the display back to pre-determined optimal contrast and brightness settings and verify the viewing conditions, such as ambient lighting, have not changed; they are then asked to repeat the login with a new code. If this process fails repeatedly the user is prompted to contact support services.
For optimal operation there is a set-up phase conducted by technical support before implementation of the verified login on a given display. These facilities are not available to general users. The first step is to determine the optimal contrast and brightness settings for the display. This is achieved using a set-up protocol [6] based around the Society of Motion Picture and Television Engineers (SMPTE) test pattern [7]. This protocol is as follows:- Set contrast to approximately 50% of maximum and the brightness control to maximum.
- Decrease brightness until the outer black region of the 0/5% test region of the SMPTE test pattern just starts to turn from grey to black without losing contrast with the inner region.
- Maximize the contrast and then start reducing it until the outer white region of the 95/100% test region of the SMPTE test pattern stops glaring but remains reasonably bright with good contrast between the inner and outer regions.
The second stage is to determine the threshold grey scale intensities for each of the four regions. This is achieved using the service mode of the verified login software. The optimum contrast and brightness settings derived from the set-up described above are entered into the software. The verified login code is then displayed with function buttons that allows the operator to step back and forth through 10 letter contrasts of 5%, 2%, 1%, 0.8%, 0.6%, 0.4%, 0.3%, 0.2%, 0.1% and 0% of maximum grey level relative to each background. The letter grey level values are added to the grey level values of the black and mid 2 backgrounds and subtracted from those of the mid 1 and white backgrounds. For each of the four backgrounds the threshold letter grey levels are determined. These display-specific threshold values and the optimum contrast and brightness settings are saved to a file on the computer.
To assess the practicability of the verified access, five observers were asked to use the image display software on three monitors for a range of contrast and brightness settings, including the optimal settings. The success rate for repeating the challenge code was recorded.
A range of displays were tested including a standard computer CRT (cathode ray tube) Monitor, a diagnostic grade CRT and a high quality flat panel monitor (see Table 1). All monitors were used with a Matrox Millenium G450 Dual Head LX video card (Matrox Graphics Inc., Dorval, Canada).
|
Results
|
|---|
The percentage correct responses for each monitor and all observers are shown in Tables 1–3
. The CRT monitor achieved between 0% and 100% successful logins for the range of contrast and brightness settings. The reporting monitor achieved between 60% and 100% and the TFT (Thin Film Transistor Liquid Crystal Display) between 20% and 80%. At the optimal settings both the CRT and the reporting monitor achieved a 100% success rate whilst the TFT only achieved 60%.
View this table:
[in this window]
[in a new window]
|
Table 2. Results for the CRT (cathode ray tube) monitor, which had the worst performance for negative settings but a 100% positive response for optimum values and above. Only the black region showed any sensitivity to contrast and brightness variation
|
|
View this table:
[in this window]
[in a new window]
|
Table 3. Results for the reporting monitor which performed the most consistently and had a greater latitude of performance resulting in more successful logins. Only the black region showed any sensitivity to contrast and brightness variation
|
|
View this table:
[in this window]
[in a new window]
|
Table 4. Results for the thin film transistor monitor showing that it was unable to provide 100% correct response in the black region. The black region and the mid 2 region are the most sensitive
|
|
|
Discussion
|
|---|
The results from the observer study showed that the verified login is most successful for detecting contrast and brightness values lower than the optimum values. They also indicate that each monitor has its own sensitivity range. The most consistent performing display was the reporting monitor. The worst performing was the TFT, as it was not possible to achieve 100% login success at the optimal settings. The most sensitive regions were the black and mid 2. The study showed no sensitivity for the other regions and ultimately they may not be required. More extensive testing of a wide range of monitors is required to confirm this.
The verified login was able to prevent access to the imaging software where display settings were set below the optimum. This is important as these settings have the greatest impact on display latitude and subsequently would have the greatest impact on image presentation. In addition, when the verified login is configured, the suitability of the display, i.e. reporting, review or non-clinical use, can be determined and the software programmed to instruct the user as to the permitted application when they log in. Launching the software would be logged and the acknowledgment of the user to the use of the system recorded. Inevitably this method of verified login will also be testing the observer's contrast sensitivity and perceptual ability; however this may unintentionally also be advantageous.
A set-up stage was used in this work to provide optimum levels of operation. Alternatively this stage could be by-passed and a default set of parameters used which would ensure a minimum level.
A restriction of the system is that the monitor needs to have digital controls on contrast and brightness levels for precise adjustment. Not all monitors have this and without it set up will be more arbitrary. A significant disadvantage is that it may be possible to defeat this system, for example by setting contrast and brightness to 20% over the optimum, although it could be argued that at these settings a minimum level of contrast is still preserved. Another limitation of this method is that the initial set-up is subjective and conducted by one observer only. It is assumed that this person will be an expert in monitor set-up and will have a good perceptual ability. However, this aspect was not assessed in this study and may have been a factor in the poor performance of the TFT display where only 60% of logins were successful at the optimum settings. The range of contrasts presented during initial set-up was discrete and may bridge the actual threshold value of the display; a continuously variable contrast could be used to overcome this. It is also possible that not all alphanumeric fonts have equal perceptual presentation. This would need further investigation and a suitable font identified before implementation. A more complete understanding of the relationship between monitor settings and diagnostic efficacy is also required before this system can be properly optimized.
|
Conclusion
|
|---|
A method for controlling access to soft copy display software and ensuring a minimum level of contrast presentation has been presented. Although this system is not foolproof, it does introduce a layer of audit and quality control that may otherwise be missing.
|
Footnotes
|
|---|
UK Patent applied for. Application Number: 0416904.1 
Received for publication September 23, 2004.
Revision received February 16, 2005.
Accepted for publication March 21, 2005.
|
References
|
|---|
- 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]
- Mertelmeier T. Why and how is soft copy reading possible in clinical practice? J Digital Imaging 1999;12:3–11.
- Ly CK. SoftCopy display quality assurance program at Texas Children's Hospital. J Digital Imaging 2002;15(Suppl. 1):33–40.
- Poynton C. Digital video and HDTV: algorithms and interfaces. San Francisco: Morgan Kaufmann Publishers, 2003.
- NEMA. Digital Imaging and Communications in Medicine (DICOM) Part 14: grayscale display function. Virginia: National Electrical Manufacturers Association, 2003.
- Jervis SE, Brettle DS. A practical approach to soft-copy display consistency for PC-based review workstations. Br J Radiol 2003;76:648–52.[Abstract/Free Full Text]
- SMPTE. RP 133: Specifications for medical diagnostic imaging test pattern for television monitors and hard-copy recording cameras. New York: Society of Motion Pictures and Television Engineers, 1991.
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]
|
|
|
Top
Abstract
Introduction
