Manual Menu INTRODUCTION TO IMAGE PROCESSING

Nearly any picture, regardless of the form (print, slide, or negative), has some problem with color balance, tonal range, contrast, color fading, or blemishes or scratches.  Slides and negatives retain more information than prints, which may have lost more than 40% of the original information in the printing process.

Some image processing examples follow, divided into reflective (prints and paper), negative transparency, and positive transparency image scans.  These example feature many processing techniques, some of which are:

• Level Processing: RGB images can be broken down into four color histograms, the original RGB image plus the separate Red, Green, and Blue images.    Often, there will be dead space at either end of the histogram, and removing this dead space often improves the image, especially for an image scanned from a print.   A marker can be moved to alter the color balance, based on quarter, half, or three-quarter tones.  Some kinds of lighting need similar correction.    Fluorescent lighting often needs a reduction in the green, and high-altitude pictures often need a reduction in the blue.
• Curve Processing: curve processing is closely related to Level Processing.  Instead of a histogram, there is a curve for the RGB and the Red, Green, and Blue images.  Curves can be used to adjust the balance of quarter, half, and three-quarter tones at the same time.  By increasing three-quarter tone intensity and decreasing quarter tone intensity, the contrast will be increased.  This method is preferred over the basic intensity and contrast change method, which tends to throw out information.
• Scratch and Defect Removal: This may be accomplished in many ways.  A Nikon 2000 film scanner can remove fingerprints and scratches during the scanning process, though it will increase the scan time.  Photoshop can remove many defects by 1) using the Filter-Noise- Dust and Scratches feature to create a blurred copy of the image, 2) using the Filter-Noise- Add Noise feature to put a little noise back in, 3) using the History-Snapshot feature to create a merge copy of the image, and then 4) using the history tool to either darken or lighten defects by exposing the blurred image underneath. 
• Patch Overlay:  Some defects can be removed by using the Box Select and Copy features.  Part of the nearby undamaged image can be copied onto the damaged part; then the Scratch and Defect Removal, described above, can be used to blend the repaired portion of the image.
• Unsharp Mask: This is the preferred method for sharpening an image, since it can specify the noise level and avoid sharpening the noise too.
• Color Correction: Some color correction can be obvious, but much of it can also be very subjective.  Colors can be altered in Level or Curve manipulation. Levels also has a feature where color balance can be altered based on the fact that there is a medium gray image somewhere in the picture.
• Color Replacement: This method can be used to correct colors, but it can also enhance or exaggerate colors.  It can alter the color hue, saturation, and lightness.  This can also be useful for adjusting paper scans (usually B&W), to lighten the background and darken any printed information.   This can be useful whether paper scans are left in graphic form or converted to a word processor format by an OCR program.

Click on the image to view print, negative, or positive processing.

	Reflective Print or Paper Processing:
	Negative Transparency Processing.
	Positive Transparency or Slide Processing.

Last modified Friday, November 01, 2002