Electronic Image Management
Working with Video images
How to capture, manipulate and publish your images.
One of the most important aspects of our PEET project is our work with images of the organisms. Our fungi are very small and almost all data must be obtained from microscope work and this information must be documented in some way. The best and easiest way is to capture images from the compound and dissecting microscopes. The easiest way to use large numbers of images is to have them in electronic form, on the computer. This demonstration shows how we capture our images with video systems, manipulate them for various output, store them for future use and use them for publication.
Capturing Video Images
Image data can be obtained in a number of ways: hard copy photographs can be scanned with a flatbed scanner, color slides or black & white or color negatives can be scanned using a slide scanner or images can be input directly to the computer via a video system or digital camera attached to the microscope. There are differences between the images obtained from a video system and those from scanners. The scanner lets you set the resolution of the acquired images directly; the video does not. The quality of images from slides or negatives acquired from a scanner is dependent on the quality of the scanner used. Images should be acquired at the resolution and the size (height, width) (given in %) needed for the final output. With video, you need to have a good high resolution video system (with 750 horizontal lines) to get good images. In our lab, we occasionally use various scanners for images but most of our images come from video systems and video imaging is what this demonstration will focus on.
Video Systems
One video system in use in our lab is a Dage DC-330 1/3" RGB color camera with a 0.31x reduction lens to obtain a larger field of view. This is mounted on an Olympus BH-2 compound microscope, equipped with bright-field and differential interference contrast optics. The camera has an adaptor to mount onto a Wild M5A stereo microscope. The video system sends the images directly to a Macintosh computer, a Power Computing PowerTower Pro 225 Mac/OS system outfitted with a Scion LG-3 Framegrabber to capture the images. The computer has 32 MB RAM and a 1 GB hard drive.
The second video camera we have is a Sony DXC-970 1/2" RGB color system with a 0.45x reduction lens to obtain a larger field of view. This is mounted on a Zeiss Axioskop equipped with bright-field, phase contrast, fluorescence and differential interference contrast optics. This video system sends images to a PowerMac 7500 outfitted with a Scion AG-7 framegrabber. This computer has 32 MB RAM and a 3 GB hard drive.
Image Capture
Images are captured from the video camera using Adobe Photoshop. The resolution for capturing images from a video camera cannot be chosen as with a scanner. The quality of resolution is dependent on the quality of the video system. Both of these cameras are high resolution/high quality video cameras that give excellent results. Video images appear on screen with a 72 dpi resolution; resolution of your final output (i.e. web pages, b&w or color plates for manuscripts) can be adjusted with the software program. (Image resolution is changed by a process that alters the dimensions of the image but leaves the file size the same. This leaves the quality of the image intact, so images remain high quality only changed in dimension. For methods, see B & W Plate demonstration under 2. Assembling the composite photos.) Obtain your raw images from the video by capturing the image with the highest magnification possible on the microscope and at the largest size available to the framegrabber (i.e. Width 640, Height 480 pixels). If you have done this, the image appears on your screen in 72 dpi and, using these framegrabbers, with the dimensions of about 9" width and 6.5" height with a size of about 900 kb. The images are manipulated and further processed in Adobe Photoshop for whatever final output you want. NIH Image is the image analysis software used in our lab for measuring structures etc.
A single set of color video images can be used for multiple output such as directly for the computer screen, as in web pages, or b&w plates for manuscript output or color slides printed to a film recorder.
A. Raw video images demonstration
Images captured from our video cameras are in color. Adjustments in contrast, color, sharpness, etc., should be made on these raw video images. They can actually be done at almost any step along the way, but VERY IMPORTANT, these adjustments should only be done once to an image. Contrast, color adjustments and sharpening degrade the quality of an image so they should be kept to a minimum.The following is an example of the procedures used in our lab to enhance raw images. Consult the Photoshop manual for information on using tools when description are not provided below.
The first thing you can do is rotate the image to the proper orientation using the Rotate feature under the Image Menu. Choose the degree of orientation and remember, this applies to the entire image. You may use the Rubber Stamp tool to correct any flaws in the image. This is one of most useful tools as it clones an area of the image and deposits it over the flaws, dirt, oil drops, etc. It can also be used to fill blank space on an image when making a composite plate. Hold down the Option key and click with the mouse to choose an area to be cloned. Then proceed to stamp out the flaws (click with the mouse).
Now I make contrast and brightness adjustments and I do this with the Adjust, Levels tools under the Image Menu (use only this tool, not Adjust Brightness/Contrast). On the histogram, set the left and right sliders [A, B] to the first actual points on the histogram where the dark lines begin. You can also move the sliders beyond the marks to enhance contrast and brightness even further. Try it and see what you get. Then adjust the mid-point slider [C] to desired lightness (I find that images usually print out darker than they first appear, so I usually lighten them up a bit). When you're adjusting levels this way, always set Preview box OFF [D]; this way your entire screen shows the effect of the adjustment you are making and by clicking on the upper bar of the Levels box [E] you can contrast this adjustment with what the original image looked like. If the Preview button is checked this feature won't work.
If your image has an unwanted color cast to it, you can go into each of the individual colors and adjust the histogram for that color until you get your desired effect (in the Levels box, look under Channel RGB, click on arrows to the right). Otherwise you can use Color Balance (see next demo).
Click OK to keep the changes. Again, remember, always make these adjustments only once. Don't go back and try to repeat the process because it only leads to degraded images. If you go back into Adjust Levels after saving the changes you just made you can see the histogram now has gaps in it. This shows that actual image data has been removed, which is what you are doing when adjusting levels.
If your image has an unwanted color cast to it, you can adjust colors under Adjust: Levels (previous demo) and go into each of the individual color histograms, changing the sliders for that color until you get your desired effect (look under Channel RGB, click on arrows to the right). Or you can useAdjust: Color Balance also under the Image Menu, again adjusting the sliders for that color until you get your desired effect. For this adjustment also, always set Preview box OFF [A]; this way your entire screen shows the effect of the adjustment you are making and by clicking on the upper bar of the Color Balance box [B] you can contrast this adjustment with what the original image looked like. If the Preview button is checked this feature won't work.
Using either method, adjusting colors can get tricky as the colors work blended with each other and are not independent. The best thing is to get the best original image with the most accurate color. I know this isn't always possible and you will end up having to do something with color eventually. (However, you don't need to make color adjustments if all you want are grayscale images.)
The next thing I do is to sharpen my image somewhat. I do this with Sharpen, Unsharp Mask under the Filter menu. Again, do this step only once because it can drastically change your image. What you want to do is only sharpen the edges within your image. Don't use the Sharpen Edges tool because it gives you no control over how much sharpening you can do. Set radius to 1.5 [A] and threshold to 10 [B]. This controls the width of the pixels that are modified at an edge and the difference in pixel values that must be present to constitute an edge. If the threshold is set to zero, everything in the image is a candidate for being an edge and getting sharpened. This can distort the picture tremendously (so set to 10 or above, experiment). Amount of sharpening can vary between 1 and 100% [C]; much over that and the image gets very distorted (Try it, and see what you get; it goes up to 500%). Keep preview selected [D] and click on the small image box [E] inside the tool to undo the sharpening inside the box and see how the original contrasts with the sharpened image. OK to save results.
Next you can crop the image if you like. Normally I don't crop raw images because I like to have the full size image available for future use. But the crop tool is available and very useful for later image manipulation, to remove unwanted or unnecessary areas of the image, to make it fit into a restricted dimension, it also makes the size of the image smaller (size of disc space, in KB), etc. I also don't change the resolution or image size of the raw images. I like to have the full size and native resolution available for future use.
That's all for the raw images. They get saved and recorded for future use. Save your raw images as Photoshop files, not tiff or pict, or compressed as jpg, at this time. You use these other formats for different outputs.
B. Black & White plate demonstration (Manuscript output)
These pages demonstrate the procedures for creating B & W plates for manuscripts from color video images. Examples given are for the journal Mycologia and are created in Adobe Photoshop 3.0 (some instructions for Photoshop 5.0 are provided).
1. Create a blank background plate on which you assemble the individual photos to make the composite:
For 2 column Mycologia plate final size of 6.75" x 9.25"
| % enlargement of original | set original width at | set length at |
| 100% no enlargement | 6.75" | 9.25" |
| 150% | 4.5" | 6.12" |
| 200% | 3.37" | 4.62" |
| 250% | 2.7" | 3.7" |
| 300% | 2.25" | 3.06" |
** We normally enlarge the final plate somewhat (usually 150%) because the individual video images at 350 dpi resolution tend to be rather small in dimension.
2. Assembling the composite photos
Compose the original plate filling the maximum width for whichever % enlargement you wish to use. Do not worry about the length except you cannot exceed maximum length given. The plate will be enlarged in the final step of the process.
Under Window menu, choose Palettes - Show Layers (in Photoshop 5.0, under Window menu: Show Layers )
2. Assembling the composite photos (continued)
The individual raw image will now appear as a layer in the B & W plate file [A]. The image has automatically changed from a color image to a B & W image. The resolution of this image has also changed automatically from a 72 dpi image to a 350 dpi image when it was placed onto the blank background plate. The image is also dramatically reduced in dimension.
3. Assembling the plate.
When all of the individual images have been placed into separate layers on the background plate, you are ready to arrange them in their proper location on the plate. Images overlap according to their position in the Layers palette: top ones lay above the ones below and will hide anything directly underneath them. Position of layers can be moved by dragging them up or down in the palette.
Click on a layer to work with the image in it [A]. Click on the move tool [V on the keyboard] and drag the image into position [B]. Repeat for all images.
After the images are roughly in place you can do advanced manipulation to complete the plate. For example: Select [using marquee tool, M on the keyboard] and cut/delete unnecessary pieces of an image; Select and rotate individual images; use Rubber Stamp tool [S on the keyboard] to fill in missing parts of images.
Save the plate and save a copy of the plate.
4. Enlarging the plate to fit Mycologia format.
Work with the copy of the plate; leave the original plate intact in case you make a mistake.
This step is called resampling: altering the dimensions and keeping the resolution the same. (When Photoshop does this it invents new pixels computed from the old ones.) Do this step only once and keep the % enlarged to a minimum because this procedure lowers the quality of the image. If you make a mistake, discard the plate and start over with a fresh copy of the original plate.
Under Image menu choose Image size...
Save file; go on to step 5 and 6 using this new enlarged file
5. Adding numbers, score lines and bar lines
This is the final step in composing the plate. Add a New Layer on the top of all the other layers [A]; work in this layer for adding numbers, score lines and bar lines to the plate (important!). Instructions and demonstration are for Photoshop 3.0 where all the text can be placed directly in the same layer. See below for Photoshop 5.0 instructions.
a. Numbers.
For Mycologia, there are restrictions for the fonts on figures: do not use Adobe Postscript fonts, TrueType fonts or system bitmap fonts. Helvetica is a good choice.
While the number is selected (highlighted) you can use the arrow to drag it into proper position. Click anywhere on the page to deselect number. Click on a new place on the plate for the next number. Type Tool options come up; type the number you wish to enter; choose OK. Repeat for all numbers.
In Photoshop 5.0, each individual piece of text (in this case each number) is placed in its own layer (labeled as 'T' in the right side of its layer box). You can leave each in its own layer but having all these layers can become cumbersome. In order to get all the text in one layer, you must:
5. Adding numbers, score lines and bar lines (continued)
Work in the lines and numbers layer, on top of all the other layers [A].
b. Score Lines.
Use Line Tool for the score lines between individual images. [N on the keyboard in Adobe Photoshop 3.0; in Adobe Photoshop 5.0, Line Tool is the alternative selection under the Pencil Tool (N)]. Double click on the Tool to get Line Tool options.
5. Adding numbers, score lines and bar lines (continued)
Work in the lines and numbers layer, on top of all the other layers [A].
c. Bar Lines
Use Line Tool for the bar lines. [N on the keyboard in Adobe Photoshop 3.0; in Adobe Photoshop 5.0, Line Tool is the alternative selection under the Pencil Tool (N)]. Double click on the Tool to get Line Tool options.
Lines are drawn in pixels on the screen. To change units to pixels, under File Menu, choose Preferences: Units (or Units and Rulers in Photoshop 5.0): under Units: choose Pixels.
To see the length of the line being drawn on the screen, under Window Menu: select Palettes: Show Info (in Photoshop 5.0, select Show Info ). When you begin to draw the line on the screen, the symbols delta X and delta Y appear below the regular Info Palette [F] (in Photoshop 5.0, they appear in the upper right quadrant of the Info Palette ). Delta X will give you the length of the horizontal line being drawn. (Watch numbers carefully, when you finish drawing the line the symbols disappear.)
For the folks in our lab, here is the length for the bar lines for the different microscopes. Remember to draw the correct bar lines for your enlargement factor.
| Olympus | optivar | 1 micron = # pixels | Original (no enlargement) | 150% enlargement | 200% enlargement | 250% enlargement | 300% enlargement |
| 10x lens | 1.25x | 0.65 pixels | 100 microns = 65 pixels | 97.5 | 130 | 162.5 | 195 |
| 20x | 1.25x | 1.31 | 100 microns = 131 pixels | 196.5 | 262 | 327.5 | 393 |
| 40x | 1.25x | 2.56 | 10 microns = 25.6 pixels | 38.4 | 51.2 | 64 | 76.8 |
| 100x | 1.25x | 6.40 | 10 microns = 64 pixels | 96 | 128 | 160 | 192 |
| Zeiss (Botany II lab) | |||||||
| 10x | 1.0x | 0.47 | 100 microns = 47 pixels | 70.5 | 94 | 117.5 | 141 |
| 1.25x | 0.56 | 100 microns = 56 pixels | 84 | 112 | 140 | 168 | |
| 1.6x | 0.72 | 100 microns = 72 pixels | 108 | 144 | 180 | 216 | |
| 2.0x | 0.92 | 100 microns = 92 pixels | 138 | 184 | 230 | 276 | |
| 40x | 1.0x | 1.88 | 10 microns = 18.8 pixels | 28.2 | 37.6 | 47 | 56.4 |
| 1.25x | 2.27 | 10 microns = 22.7 pixels | 34 | 45.4 | 56.7 | 68.1 | |
| 1.6x | 2.89 | 10 microns = 28.9 pixels | 43.4 | 57.8 | 72.3 | 86.7 | |
| 2.0x | 3.64 | 10 microns = 36.4 pixels | 54.6 | 72.8 | 91 | 109.2 | |
| 63x | 1.0x | 3.00 | 10 microns = 30 pixels | 45 | 60 | 75 | 90 |
| 1.25x | 3.55 | 10 microns = 35.5 pixels | 53.3 | 71 | 88.8 | 106.5 | |
| 1.6x | 4.55 | 10 microns = 45.5 pixels | 68.3 | 91 | 113.8 | 136.5 | |
| 2.0x | 5.75 | 10 microns = 57.5 pixels | 86.3 | 115 | 143.8 | 172.5 | |
| Zeiss (Sabine's office) | |||||||
| 10x | 1.0x | 0.4 | 100 microns = 40 pixels | 60 | 80 | 100 | 120 |
| 1.25x | 0.52 | 100 microns = 52 pixels | 78 | 104 | 130 | 156 | |
| 1.6x | 0.67 | 100 microns = 67 pixels | 100.5 | 134 | 167.5 | 201 | |
| 20x | 1.0x | 0.89 | 100 microns = 89 pixels | 133.5 | 178 | 222.5 | 267 |
| 1.25x | 1.12 | 100 microns = 112 pixels | 168 | 224 | 280 | 336 | |
| 1.6x | 1.43 | 100 microns = 143 pixels | 214.5 | 286 | 357.5 | 429 | |
| 40x | 1.0x | 1.8 | 10 microns = 18 pixels | 27 | 36 | 45 | 54 |
| 1.25x | 2.28 | 10 microns = 22.8 pixels | 34.2 | 45.6 | 57 | 68.4 | |
| 1.6x | 2.91 | 10 microns = 29.1 pixels | 43.65 | 58.2 | 72.75 | 87.3 | |
| 100x | 1.0x | 4.45 | 10 microns = 44.5 pixels | 66.75 | 89 | 111.25 | 133.5 |
| 1.25x | 5.65 | 10 microns = 56.5 pixels | 84.75 | 113 | 141.25 | 169.5 | |
| 1.6x | 7.3 | 10 microns = 73 pixels | 109.5 | 146 | 182.5 | 219 |
6. Save a copy [A] of the file as TIFF format [B]; this flattens all the layers [C] into a single one and reduces file size.
Check off boxes:
7. TIFF files are ready to send on diskette or ZIP disk to Mycologia; use StuffIt compression if you wish.
C. Color plates for Web pages (Screen output)
Color plates can be created for Web pages using the methods for the B & W plates. For web images I like to set rulers for pixels to better relate to screen output which is measured in pixels. Do this under File Menu, Preferences, Units. Resolution for web page images is 72 dpi, the same resolution of the raw video images, so that doesn't have to be changed.
Note: If you have images from slides that have been scanned in at a higher resolution, you should change the resolution to 72 dpi before you use them. Do this by :
1. First, resample the image to fit the screen size you want in pixels (for example, 640 x480 or 800 x 600 to fit the entire screen).
Under Image menu choose Image size...
2. Second, change the image resolution to 72 dpi.
Again under Image menu choose Image size...
Create a blank background plate in color at 72 dpi resolution. Assemble and compose the plate using the methods for B & W plates. The dimensions the plate appears on the screen at 1:1 (or 100%) are the dimensions it will have in the Web page. Save the file as jpg document with high, medium or low image quality. This reduces the file size tremendously. An example is an image that starts as 870 kb goes down to 96 kb after high jpg compression.
If you are using thumbnails in the web page text to get access to the larger screen images, you can save a separate copy of the image as a thumbnail, reduced in dimension (scaled down using Image Size under the Image Menu; for example 20%) and also reduced in file size (for example down to 18 kb). This image gets placed as the thumbnail.
D. Photoshop slides
You can also make photographic slides from your Photoshop images. This time work with rulers set to inches (File Menu, Preferences, Units). Open a new file with the dimensions 7.5" x 5", landscape or portrait with 350 dpi resolution. At this dimension the entire slide will be filled. The images need to be higher resolution than those used for web images. I've used 350 dpi resolution with good quality results but you can go higher.
Assemble and compose the plate using the methods for B & W plates.
Note: If you find that an individual image is too small for your slide, delete its layer, go back to its raw image and enlarge the image dimensions by resampling (under Image Size in the Image Menu; Select [X] Constrain: Proportions; (In Adobe Photoshop 3.0) Deselect [ ] Constrain: File size (Important!). (In Adobe Photoshop 5.0: Select [X] Resample Image: choose Bicubic); Choose % enlargement in the width or height; resolution remains at 350 pixels/inch). With this new enlarged image, then proceed to drag layer and compose plate as before.
If you wish, leave space between the individual photos. Fill excess white background with color. Use the eyedropper to choose a dark color from one of the images and then, working in the background layer fill it using the paint bucket. The result is a composite image slide on a dark background.
Save your slides as tiff files and send them to a film recorder. At 350 dpi resolution each slide takes about 3 minutes to print to the film recorder (longer if the file size is very large). You can also create these composite slides in a presentation software like PowerPoint but the resolution may not be as sharp as what you get with Photoshop.
E. Powerpoint slides
If you want to do an on-screen presentation from slide presentation software such as Microsoft Powerpoint, you can use your images as well. For Macintosh, the images have to be pict files with 16 bit color or Powerpoint will not recognize them or be able to display them properly. Resolution should be higher than 72 dpi because they will be projected on a large screen. 72 dpi is okay for computer screen presentations but not for projected presentations. We use pictures saved at 350 dpi as pict files. Insert them in the slide. Size them, add text, graphics, etc. These can be used for on screen presentations or sent to a film recorder.
Image Cataloging
To keep track of all the images, we currently use Adobe Fetch, an image cataloging program that is very easy to use (now goes under the name of Extensis Portfolio). In the future we will be exploring another program, Canto Cumulus.
The catalogs can be viewed by thumbnails, by ordered lists of the image names, file folders, etc. You can search for individual or groups of images by using key words. It tells you where the image is located and other pertinent information that you can enter. It is very easy to add and update entries. For each individual CD, I record on the CD, its own individual catalog. We also have a master catalog of all the CDs recorded. This way we can look in the master catalog and find on what CD any image is located. The only thing to remember is that you have to uniquely name each CD and you do this at the time you record it. You should be very consistant about naming your folders and files in order to best search for images or image data.
Image Storage
If you are going to deal with large amounts of image data, I strongly encourage using a CD recorder. CDs hold about 625 MB, the drives cost about $600, and the disks are very reasonabley priced ($2-$3 each). You can make several backup CDs very inexpensively. For us, because we are generating a lot of data, CD recording was the best way to go.
Other methods of storage include Zip and Jaz disks. Zip disks hold 100 MB of data, not a lot of space when you're talking about image files. The Zip drive costs about $200 and the disks are relatively expensive (about $10-$12 each). A Jaz disk holds more (1 GB of data), the drive costs about $300 and the disks run $90-125.
Publishing
Scientific journals
Creating black & white plates for publishing in scientific journals is easy to do, but you need to find out several things before you start. Does your journal and its publishing house accept illustrations in electronic form on disk? If so, at what line screen is the journal published? If it doesn't accept illustrations in electronic form on disk, you can still create your plates in Photoshop, only the final product you send with the manuscript will be different (see Film below). Most journals that publish through Allen Press should accept illustrations in electronic form on disk since Allen Press does. When we called Allen Press, who publishes our journal Mycologia, their art director gave us the guidelines for making the plates. Mycologia is printed at 175 line screen so we need to create my plates at 350 dpi resolution. Other journals print at a lower line screen and so the resolution of images should be lower. You should always match your resolution with that of the printer output. If you send a file at too high resolution to a printer that doesn't print at that high a resolution, it will spend a lot of time determining which pixel information it should discard and this will degrade the final image. There are restrictions for the fonts on figures: do not use Adobe Postscript fonts, TrueType fonts or system bitmap fonts. Helvetica is a good choice. The images should be saved in the TIFF format with LZW compression selected.
Film
If your journal does not accept illustrations in electronic form on disk, you still have one option. The final plates on disk can be brought to your local service bureau (typographer, digital imaging company, [not Kinkos, copy services don't have the high quality printers needed]) and have your plate printed out on film. They will use a high resolution printer (e.g. 3048 dpi), but be sure to specify line screen. The film can be sent with your manuscript as the originals for the plate. Any publisher should be able to print directly from the film. These films will run $20-$25 each depending on the size (8.5 x 11 to 11 x 17).
Posters
Video images can also be used for creating posters in Adobe Illustrator. A digital imaging service bureau will print it on a large 36" wide, color, 300 dpi inkjet printer.
This page is a product of the NSF-PEET (Partnerships for Enhancing Expertise in Taxonomy) grant to the Field Museum of Natural History (DEB-9521926: Studies in the Lasiosphaeriaceae. Monographs of two key genera and a family-level phylogeny).
This web page was coordinated by Sabine M. Huhndorf, Field Museum, Department of Botany