Images in the real world

The 'real 3D world' exhibits continuous gradients of time, light intensity, color, and detail, referred to as 'analog' information. Computers, by virtue of the way microprocessors (CPU) works (0-1; ON-OFF; True-False) are 'digital.' We represent the world for analytical and enjoyment purposes as 2-dimensional (2D) representations (drawings, paintings, photos) that achieved a desired result (archival and/or information transfer) but only simulated the real world. For analytical purposes this image must reproduce with fidelity the original subject. Keep in mind however, that an image is a representation of the original subject and contains biases or in some way subjectively modifies the original object.

For analytical purposes an image does not have to be an exact representation of the original. In fact, it cannot be.

• It may only approximate the original subject in detail (spatial, grayscale, color.) These parameters may be reduced without appreciably modifying the original subject, yet still maintain adequate information.
  
• The image may be a modified representation of the subject for the purposes of enhancing some parameter of the original. The resulting processed image can be used to extract information more easily.

Ansel Adams' photographs are good examples of this concept. Adams' photos are subjective representations of the real world designed by the artist to convey to the viewer a feeling of the majesty of nature. They are not snapshot photos and do not render the image gray values exactly as they were when the photograph was taken. In image processing terms, Adams modified the grayscale values of his images of nature to produce something that is not found in reality, but that is powerful in its presentation of the subject.

We take photographs of scientific subjects for archival and informational purposes that also are representations of the original object. Prior to computer image processing (and to a limited extent today) photographic processing enhanced characteristics of the subject in order to extract information or convey a more lucid representation. In black & white (grayscale) photography lens filters can be used to enhance colors (clouds, the sky, IR reflection) so that the photo shows more or reveals invisible detail. During photographic processing the range of grayscale values of the negative can be increased or decreased to include the full range from white to black (the Adams/White method). Whatever the manipulation, images are only representations of the real world.

Image processing has changed the way we manipulate images. These days computers are the medium through which images are modified and analyzed. They have opened to everyone the ability to create and/or modify images for conveying or extracting information. The first step in the process is getting the image into the computer. This is called image capture.

Resolution

• Microscopy resolution is defined as the minimum distance between two points or lines where they can be distinguished. Resolution is determined by the wavelength of the illuminating light and the aperture of the collecting lens (numerical aperture.) Using UV illumination and the best objective the maximum resolution is 0.24um.
  
• Video resolution is expressed as the total number of black and white lines displayed relative to the size of the active picture area (usually height.) Due to edge effects, the realized video resolution equals the number of active scan lines X 0.7 (the 'Kell Factor'). For NTSC this is 486 X 0.7 = 340 resolved horizontal lines. The resolution in a NTSC video signal is quite poor.
  
• Computer image spatial resolution is measured (usually) in dots per inch (dpi) or pixels per inch.
  
• Grayscale resolution is a binary measure of the number of possible gray values. Most image processing and image capture programs express grayscale in 8-bit format or 2e8 (256) levels of gray. For comparison, our eyes can distinguish only about 32 levels of gray. Another term for grayscale resolution is bit-depth.
  
• Color resolution on a computer monitor is usually a binary measure of 256 (2e8) possible values of each of R, G, and B. This results is a total of over 16 million colors (256x256x256), and is referred to as 24-bit color.
  
• Printing resolution is determined by the halftone screen frequency measured in lines per inch (lpi). This is a reflection of the size of the small, solid color dots in any printed image. A typical halftone screen frequency for newspapers is 65-85 lpi. Magazines print at 133 lpi, 600 dpi laserprinters at 85 lpi, and high-quality publications print at 150-175 lpi.