AJR Get Involved! Join ARRS Today
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
 QUICK SEARCH:   [advanced]


     


This Article
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Taylor, G. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Taylor, G. A.
Social Bookmarking
 Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?
AJR 2003; 180:329-331
© American Roentgen Ray Society


Photoshop for Radiologists

Improving Image Contrast

George A. Taylor1

1 Department of Radiology, Children's Hospital, 300 Longwood Ave., Boston, MA 02115.

Received July 10, 2002; accepted after revision July 22, 2002.

 
Address correspondence to G. A. Taylor.


Introduction
Top
Introduction
Contrast Improvement Tools
Conclusion
References
 
This is the second in a series of articles designed to help the radiologist with the process of converting 35-mm teaching slides to electronic files for insertion into PowerPoint (Microsoft, Redmond, WA). The first article [1] covered several basic steps in the initial process, including conversion of multichannel RGB files to gray-scale images, removal of patient identification, repositioning of text, and image resizing. This article presents some ways to improve the contrast and dynamic range of the image after those basic steps have been performed.


Contrast Improvement Tools
Top
Introduction
Contrast Improvement Tools
Conclusion
References
 
Photoshop (Adobe Systems, Mountain View, CA) provides two general methods for controlling the intrinsic contrast of an image. Both the "Levels" and the "Curves" tools allow a great deal of control over the distribution of pixel densities and how those densities are displayed in the final image output.

Levels
Two "Levels" tools—"Levels" and "Auto Levels"—may be accessed via the pull-down menu under "Image." After selecting "Image," select "Adjust," and then select either "Levels" or "Auto Levels." (Fig. 1A).



View larger version (89K):
[in this window]
[in a new window]
[as a PowerPoint slide]
 
Fig. 1A. Coronal sonogram of male neonate with postinfectious hydrocephalus. Low-contrast gray-scale sonographic image is shown on right. "Levels" tool is selected by clicking "Image" menu bar, then "Adjust," and then "Levels," as shown in upper left corner. Histogram located in lower left of figure shows distribution of pixel densities in sonographic image. Highest (brightest) pixel value is depicted by light gray arrowhead to right of histogram, lowest (darkest) value is depicted by dark gray arrowhead to left of histogram, and middle gray is depicted by medium gray arrowhead near center of histogram. Outermost markers are located beyond the brightest and darkest pixels in image.

 

The "Levels" dialogue box consists of a histogram showing the relative distribution of density values in the image. It allows the user to change the dynamic range of the image analogous to window and level controls on an imaging workstation. The outermost arrowhead markers denote the window width of the values displayed, and the center marker represents the midpoint of densities displayed (window level). The highest (brightest) pixel value depicted is controlled by the light gray arrowhead to the right of the histogram, and the lowest (darkest) value is controlled by the dark gray arrowhead to the left of the histogram. The effect of "Levels" correction can be seen in Figures 1A and 1B. Figure 1A shows the relatively low image contrast present in the original scanned image and its accompanying histogram. Note that the outermost markers are at a distance from the brightest and darkest pixels in the image, and the midpoint level is set slightly to the right of the histogram center. By bringing the outermost arrowhead markers to the edges of the histogram and the midpoint marker closer to the center of the histogram, the image contrast is markedly improved (Fig. 1B).



View larger version (79K):
[in this window]
[in a new window]
[as a PowerPoint slide]
 
Fig. 1B. Coronal sonogram of male neonate with postinfectious hydrocephalus. After modification using "Levels" tool, image shows relocation of outermost arrowhead markers to edges of histogram and midpoint marker to left of center of histogram, toward denser pixels. Note improved image contrast. Arrows show reverberation artifact in third ventricle.

 

"Auto Levels" is, as the name suggests, an automatic image-correction tool that is useful for quick and relatively straightforward corrections in contrast. The "Auto" function essentially makes the darkest pixels in the image become black and the brightest pixel become white, and then it spreads out the intervening density values evenly between them [2]. The effect of "Auto Levels" correction can be seen in Figure 1C. This figure is similar to Figure 1B, except that Figure 1C does not depict the reverberation artifact in the dilated third ventricle as well. The difference lies in the location of the midpoint arrowhead marker on the histogram. In Figure 1B, the midpoint was arbitrarily placed toward the denser pixels, whereas "Auto Levels" placed the midpoint in the center of the histogram.



View larger version (76K):
[in this window]
[in a new window]
[as a PowerPoint slide]
 
Fig. 1C. Coronal sonogram of male neonate with postinfectious hydrocephalus. After modification using "Auto Levels" tool, image shows relocation of outermost arrowhead markers to edges of histogram and midpoint marker placed in center of histogram. Figure C is similar to B except that it does not depict reverberation artifact in dilated third ventricle as well.

 

In general, I begin image contrast correction with "Auto Levels" and fine-tune the image with the "Levels" tool as needed.

Curves
The "Curves" tool may be accessed via the pull-down menu under "Image." After selecting "Image," select "Adjust," and then go to "Curves." The graphic interface is based on an x-y axis. The y axis represents the range of density values present in the image, and the x axis represents how each of those values is displayed as an output [3]. In the case of electronic images, the graphic interface reflects the densities displayed on the monitor. The default curve shows a straight linear correlation between density values and their display. In gray-scale mode, darker pixels are represented by the upper part of the curve and lighter pixels by the lower part. By shaping the curve, one can affect a specific range of density values without affecting others.

The first step in using the "Curves" tool is to sample the image to determine which pixel densities need to be changed. Select the "Eye Dropper" tool from the tool palette (Fig. 1D), place it over a light or dark area of the image, and click. The density of the selected pixel will be shown as a small circle along the curve. In Figure 1D, the "Eye Dropper" was placed over a relatively dark area, and the pixel density is represented by a circle on the upper third of the curve. The curve can be reshaped by clicking on a portion of the curve and dragging it up or down, depending on the change required. Figure 1E shows a reshaped curve and its effect on the image. The curve now has a sigmoid shape, in which the upper curve (representing darker pixels) is shifted upward, and the lower curve (representing lighter pixels) is shifted downward relative to the original curve. Note that the mid portion of the curve (representing middle gray pixels) has not been affected. The resulting image contrast is much closer to that of the analog original.



View larger version (91K):
[in this window]
[in a new window]
[as a PowerPoint slide]
 
Fig. 1D. Coronal sonogram of male neonate with postinfectious hydrocephalus. Low-contrast gray-scale sonographic image (same as A) before contrast modification. Curves tool is selected by clicking the "Image" menu bar, then "Adjust," and then "Curves," as shown in upper left corner. Default "Curves" graphic interface shows straight-line relationship between range of density values present in image (y-axis) and how each of those values is displayed as output (x-axis). "Eye Dropper" tool was selected from tools palette (white arrow) and placed over a relatively dark area of sonographic image. Density of image pixel selected is represented by circle on upper third of curve (black arrow).

 


View larger version (75K):
[in this window]
[in a new window]
[as a PowerPoint slide]
 
Fig. 1E. Coronal sonogram of male neonate with postinfectious hydrocephalus. After modification using "Curves" tool, image shows curve with sigmoid shape: darker pixels (upper curve) are shifted upward, and lighter pixels (lower curve) are shifted downward. Note that mid portion of curve (middle gray pixels) remains straight. Resulting image contrast is close to that of analog original.

 

When modifying a curve, it is important that the changes be smooth and gradual. A jagged or sharply angled curve will result in a degraded, unacceptable image.

Other Controls
Two other contrast adjustment tools are available under "Adjust" in the "Image" menu (Fig. 1D): "Brightness/Contrast" and "Auto Contrast." Both may be used as a quick fix for contrast problems. However, I rarely use them because they tend to overadjust the image. "Levels" and "Curves" give the user much finer control over the final product.


Conclusion
Top
Introduction
Contrast Improvement Tools
Conclusion
References
 
The tools described above are relatively easy to use and will result in image contrast that is much closer to the slide or analog print used as the original for the digital file. It is important to avoid the temptation to over-correct the image: the goal is to return the scanned image to its original appearance.

The next article in this series will cover methods of removing dust and scratches from the scanned image.


References
Top
Introduction
Contrast Improvement Tools
Conclusion
References
 

  1. Taylor GA. Photoshop for radiologists: initial steps in image preparation. AJR 2002:179:1411 -1413[Free Full Text]
  2. Ang T. Contrast and levels. In: Ang T. Silver pixels: an introduction to the digital darkroom. New York: Amphoto Books, Watson-Guptill Publications, 1999:30 -35
  3. Caponigro JP. (R)evolution of the zone system. Photo>Electronic Imaging 2001;9:20 -25

Add to CiteULike CiteULike   Add to Complore Complore   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us   Add to Digg Digg   Add to Reddit Reddit   Add to Technorati Technorati    What's this?


This article has been cited by other articles:


Home page
Am. J. Roentgenol.Home page
C.-S. Yam, J. Kruskal, and M. Larson
Using Movie Clips in PowerPoint Presentations: Part 2, Movie Editing
Am. J. Roentgenol., November 1, 2005; 185(5): 1371 - 1376.
[Abstract] [Full Text] [PDF]


Home page
Am. J. Roentgenol.Home page
C. Sistrom and J. Honeyman-Buck
A Simple Method for Importing Multiple Image Files into PowerPoint
Am. J. Roentgenol., June 1, 2004; 182(6): 1591 - 1596.
[Abstract] [Full Text] [PDF]


Home page
RadioGraphicsHome page
M. A. Haider
Extending PowerPoint with DICOM Image Support
RadioGraphics, November 1, 2003; 23(6): 1683 - 1687.
[Abstract] [Full Text] [PDF]


Home page
Am. J. Roentgenol.Home page
G. A. Taylor
Color Correction and Automating Repetitive Tasks
Am. J. Roentgenol., August 1, 2003; 181(2): 383 - 386.
[Full Text] [PDF]


Home page
Am. J. Roentgenol.Home page
G. A. Taylor
Sharpening the Image
Am. J. Roentgenol., July 1, 2003; 181(1): 43 - 45.
[Full Text] [PDF]


Home page
Am. J. Roentgenol.Home page
G. A. Taylor
Removal of Dust and Scratches from Electronic Image Files
Am. J. Roentgenol., June 1, 2003; 180(6): 1533 - 1536.
[Full Text] [PDF]


This Article
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Taylor, G. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Taylor, G. A.
Social Bookmarking
 Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?


HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS