Print
Email
Practical Considerations When Altering Digital Images
Publication Name:
Current Protocols Essential Laboratory Techniques
Unit Number:
Appendix 3B
DOI:
10.1002/9780470089941.eta03bs00
Online Posting Date:
October, 2008 Abstract
Worldwide, Adobe Photoshop is the primary means in research for altering images. Training on the use of this powerful tool most often is peer to peer within a laboratory, or it is self-taught through trial and error. As a result, varying amounts of visual data can be removed unwittingly; contrast- and brightness-enhancement of images can be done without adjusting color and grayscale levels to conform to the output (which includes publication, projection, web placement, and inclusion into Microsoft Word and Adobe Acrobat); images can be unethically altered; and pixelation can be introduced. Correct methods for use of Adobe Photoshop are outlined in this unit so that these errors can be avoided.
Keywords: pixelation; sampling resolution; brightness; contrast; saturation; pixels; resampling; gamma; midtones
Table of Contents
- Introduction
- Sampling Resolution
- Resampling
- Acquiring Images
- Photoshop and Scientific Image–Analysis Programs
- Optimizing the Display
- Using Images from Vector Programs and PowerPoint
- Altering Images Using Photoshop
- Inserting Files into PowerPoint
- Literature Cited
- Figures
- Tables
Figures
-
Figure A.3B.1“Stairstepped” lettering (aliased; A) and anti-aliased lettering (B). Note ghosting of aliased sides of lettering. -
Figure A.3B.2Noncompressed image (A) and JPEG compressed image (B). Compression was done in Photoshop at a low setting (for higher compression). Image in B was sharpened to aid in showing JPEG compression artifacts. -
Figure A.3B.3Arrows point to clipped or oversaturated parts of the image. Note loss of detail in white areas, and creeping of black ink into features in the clipped, black background portion of the image. -
Figure A.3B.4Areas of primary colors for light are marked red, green, and blue. Note the excessive lack of overlapped gamut in the green hues. Note, also, the variable overlap between what is displayed on a computer screen and what hues can be printed on a printing press in the blue and red colors. -
Figure A.3B.5Image in A shows area of a single image in which no color fringing occurs. Image in B shows color fringing at edges of cells. These edges were colored cyan at the upper, left area of each cell, and a purplish hue at the lower, right of each cell. Color fringing artifacts disappeared when image was made from a RGB Color image to Grayscale. Fringed areas were amplified and made into distinct tones to show the effect. -
Figure A.3B.6Images show two dialog boxes that appear when creating PDF documents on Windows with Adobe Acrobat 7 Professional. Note that the document is created by printing, not by saving or exporting. Click on “Properties” to open tab for “Adobe PDF Settings,” and then choose “Press Quality” from the drop down list for best reproduction of both lettering and images. -
Figure A.3B.7Recommended timeline for global image manipulations. Manipulations shown in bold are described in this appendix; other manipulations are beyond its scope. -
Figure A.3B.8Two examples are shown of information presented in the image window in Photoshop. The top image shows that it is open at 33%, a reference to the size of the image in relation to a legacy Apple computer screen. It is not a reference to the dimensions of the image in inches or centimeters. The top image window bar also shows the Mode of the image, in this instance an RGB Color image at 8-bits per channel. The bottom bar shows a Mode of Index#, indicating that the image contains 256 levels of gray or of colors. That mode of image can be altered with full functionality of Photoshop after conversion to RGB Color or to Grayscale. -
Figure A.3B.9A completely black image contains detail, as shown by the histogram in the histogram dialog box (arrow). In this instance, it is an image converted from 12-bits to 16-bits in software that came with the camera. The 16-bit image was opened in Photoshop CS2 on Windows, only to contain an 8-bit readout of grayscale values, though the image Mode is yet considered to be 16-bits. -
Figure A.3B.10Image in A has been converted from 16-bits to 8-bits using the conversion option under Image>Mode>8-bits/channel. Then the image was brought back to values that existed in the original display using Levels. Note extreme loss of gradients in the gray values: instead, a topographical look containing limited and separate values results. Image in B was corrected in Levels first, and then converted from 16-bit/channel to 8-bit/channel using Mode. Note the absence of banding. -
Figure A.3B.11The Palette Options dialog box can be accessed from the Info Box by clicking on the top, right arrowhead to reveal a drop down box. To guarantee a display of RGB values (when the image is an RGB Color image), that is chosen as the first readout. The actual color is chosen for the second readout. Pixels are chosen for the unit. Note that Grayscale images will show only the percentages of black ink used when publishing to paper output. This is shown as the K: value. -
Figure A.3B.12The Color Sampler tool is revealed from the toolbar by clicking and holding with the mouse on the eyedropper icon. Set the Sample Size to 3 × 3 or 5 × 5 from the submenu (in versions later than Photoshop 5.0) so that pixels surrounding areas subsequently chosen will be averaged together to obtain values. Once the brightest and darkest significant areas are chosen by clicking on the image with the Color Sampler tool, the numbered values appear at the bottom of the Info dialog box. -
Figure A.3B.13When layers are created on an image, the topmost layer “blocks” the view of other layers when the layer mode is set to Normal (the default setting). Varying levels of transparency can be obtained by clicking on the arrowhead adjacent to “Normal” and then choosing from the drop down list. In this instance, “Hard Light” is chosen. The opacity of the layer can be subsequently altered using the Opacity slider, in this instance to 50%. -
Figure A.3B.14Limits to maximum white and minimum black values can be indicated in the Levels dialog box. In this instance, the white limit is being set by double clicking on the white eyedropper tool and setting the red, green and blue values to 240. That setting is generally within the range of most printing devices. -
Figure A.3B.15The white eyedropper tool can be chosen in the Levels dialog box, and then a white portion of the image can be clicked to auto-correct color. Several white areas may need to be clicked on until the image is fully corrected. Evaluation is done by eye. -
Figure A.3B.16Image in A was blue colorized in confocal software with a standard blue LUT (Look Up Table) setting, then made into grayscale. Image in B shows solid gray toned areas that will not reproduce accurately. Note that all features have the overlay, indicating the lack of overlapped gamut values between a computer screen display and a printing press output. -
Figure A.3B.17Setting limits for fluorescence can be accomplished by first of all sampling the brightest significant whites and blacks. These can be adjusted in Levels while viewing respective numeric readouts in the Info box. In this instance, a blue colorized image shows values only in the blue channel. -
Figure A.3B.18To conform displayed images to the gamut of a printing press, the Hue/Saturation tool can be used to desaturate offending colors. In this instance, a blue-colorized image is desaturated by sliding the Saturation slider to the left to a value of –35. Arrowheads show gamut overlay tones that still exist, showing likely areas of poor reproduction. These areas are ignored because of past experience when publishing, opting, instead, for a desaturation setting of approximately –32. -
Figure A.3B.19Layers can contain the alterations that are the outcome of using Levels sliders. Rather than select Levels from the Image>Adjust(ment) drop down list, Levels can be found by choosing the circle icon in the Layers dialog box. When that selection is made, the alterations are applied to a layer that can be eliminated or edited later on without affecting visual data. -
Figure A.3B.20The Image Size dialog box can be divided into the inherent resolution readout (in pixels) and the resolution were it to be outputted. In this instance, the image is read in a curious way: the resolution is indicated as 1 with a width and height that matches pixels across and down. If this image is to be printed or prepared from PowerPoint with its inherent resolution maintained, Resample Image should be left unchecked. Simply set the width and height and ignore resolution. The output resolution only needs to be changed (with Resample Image checked) when outputting to specified resolutions, such as those given by publishers. Resampling may also need to be done when original resolution is low. -
Figure A.3B.21Once an image is placed on a 300 pixel/dot per inch page for inclusion in a figure (or plate), then individual images can be rescaled in dimension to match sizes at desired output dimensions. Choose Scale under Edit>Transform and hold down the shift key when resizing to keep width to height ratios the same. If the Info box is open, then the percent scaling can be noted and repeated for subsequent images. -
Figure A.3B.22From left to right, the grid is revealed over the image when clicking on it from Window>View. The spacing of the grid can be changed by choosing Guides, Grid & Slices from Edit>Preferences (or Photoshop>Preferences on a Macintosh). The “Guides” dialog box allows users to input values for grid sizes. -
Figure A.3B.23If images included in PowerPoint are derived from Photoshop, the document size is set with the default PowerPoint window size in mind: 10 in. in width, 7.5 in. in height. In this instance, the width is set to 9 in. to almost fill the PowerPoint slide window with the Resolution left at 153.333 because it is yet above poster printing resolutions of 150 dots/pixels per inch. Resample image is left unchecked so that the resolution at a particular output can be determined before making a decision to upsample, when necessary.
Literature Cited
| Literature Cited | |
| Leong, F.J., Brady, M., and McGee, J.O. 2003. Correction of uneven illumination (vignetting) in digital microscopy images. J. Clin. Pathol. 56:619-621. | |
| Margulis, D. 2006. Professional Photoshop: The Classic Guide to Color Correction, 5th ed. Peachpit Press, Berkeley, CA. | |
Editorial announcements
Troubleshooting Tips
|
TOOLS & CALCULATORS |
Join the Conversation
Post new comment