Color Images: Generation and Interpretation

Iqbal S. Sandhu, M.D.

 

Introduction

The topic of discussion in the previous column was the generation and storage of a digital image based on the sampling of an analog phenomenon, i.e., the computer INPUT. Conventional monochrome output was assumed, i.e., the absence of a signal is represented by black, a strong signal (high amplitude) by white and intermediate strength signals by shades of gray. We assumed that the mapping would be linear in the digital domain and logarithmic in the sound domain (decibels). In this issue, we concentrate on the visual display or computer OUTPUT, with emphasis on color display and noting that shades of gray are specific shades of color.

Figure 1

The image OUTPUT is the mapping or representation of a number at each point in an image (pixel) to visible light intended for our eye. The range of numbers stored in an image matrix is typically limited to 2 to the 8th, 2 to the 16th, and 2 to the 24th. The output or image display is produced by coupling each of the possible numbers in a digital image to a voltage applied to the acceleration grid of a single gun Cathode Ray Tube (CRT). A conventional approach will map the number 0 in an 8-bit image to 0 volts on a CRT and the number 255 to the voltage required for maximum practical brightness on the CRT. However, there is generally no limitation on how this mapping is done, and we might instead suppress low signal strength ultrasound noise by mapping many small numbers, instead of just 0, to black.

Figure 2

If we use a three-gun (Red, Green, Blue) color CRT instead of a single-gun grayscale model, we can map the value of a pixel to all three guns equally and produce the same grayscale image produced by the single gun system. Or, we could be more creative and map numbers from the image matrix to different intensities on the Red, Green and Blue guns to produce a rainbow of colors to dazzle the eye and justify the machine to the administration. This kind of mapping has typically been met with enthusiasm similar to the colorization of Ansel Adams' masterpieces. Color mapping without a sound physical basis is generally viewed as extraneous noise confusing the interpretation of an image. Apart from aesthetic considerations, and potential distraction, use of color can in fact be deceptive. When we map a grayscale image, we often do it linearly. When color is used to display continuous data, mapping becomes an issue. At some point along the continuum of data, increasing the brightness of one color (representing one physical parameter) transitions to display a different color, implying a second physical parameter or a large change in intensity value i.e., a "border." At what point along the continuum of data does red change to blue, and what "borders" do these two colors represent? It is easy to see how this could be confusing and used incorrectly.

However, one use of color that has significant clinical value is the representation of a second parameter, i.e., fluid flow, with overlaid on a traditional grayscale representation of acoustic impedance. Fluid flow is measured by its Doppler Effect or the change in frequency of a ultrasound wave reflected from a moving object. The Doppler ultrasound image is a separate image from the static echo ultrasound image. A Power Doppler monochrome image, usually orange, is combined with the acoustic impedance image, usually gray, for display. This provides the visualization of two physical parameters (acoustic impedance and fluid flow) in one image. If the direction of the fluid flow were important, separate colors could be assigned to blood flow toward or away from the transducer. Being able to distinguish blood vessels from ductal and cystic structures is valuable during ultrasound directed aspirations of lesions as it helps the physician avoid puncturing blood vessels and minimizes bleeding complications.

References

1. Moving Theory into Practice, Digital Imaging Tutorial http://www.library.cornell.edu/preservation/tutorial/ Cornell University Library.

2. Kruenen Ben. Big Ben?s Digital Imaging Tutorial http://www.bigbenpublishing.com.au/digital/

3. Bouton, DG, Bouton, B, Kubicek, G, Nathanson, MZ, Rich J, Ward A. Inside Adobe Photoshop 6, Limited Edition, 2001, Bouton, DG (editor) New Riders Publishing, Indianapolis, IN.

 

 




Editorial Board:
Manoop S. Bhutani, M.D.
Galveston, TX
William R. Brugge, M.D.
Boston, MA
Peter R. McNally, D.O.
Denver, CO
Iqbal S. Sandhu, M.D.
Salt Lake City, UT
Thomas J. Savides, M.D.
San Diego, CA

Copyright © 2002, University of Colorado, All Rights Reserved
Privacy Policy