Manual Menu See Introduction to Image Processing

This page is about digital image processing.  Specifically, it is about processing images scanned from from photographic prints, slides, and negatives, and from printed paper. Methods described here are also applicable to images produced with the newer digital cameras, but none of the images on this website are from a digital camera.   Digital cameras are very expensive (for higher resolution), are still being developed, and employ a multitude of (ever increasing) storage formats, which must be downloaded to a computer.  For information on digital cameras, see Smile, Digital Cameras, for a more extensive discussion.

Although it stands alone, this page is used in conjunction with a new digital imaging service for users of photographic film.   Digital image scanning and image processing is one of the most promising new internet businesses.  Despite the newness, there have already been many casualties in this field.  If the site, LIST of low cost scanning services, is examined, it will be seen that almost half the referenced URL's no longer exist, or are no longer in the scanning business.  Here is a list of some current scanning services:

There are at least four types of scanners:

1. Flatbed scanners, which can can prints and text (reflective media),
2. Film scanners, which can scan negatives and slides,
3. Combination flatbeds, which can scan negatives and slides as well as reflective media, and
4. Drum scanners, which can scan reflective media.

Drum scanners are usually used by newspapers and publishing houses for pre-press scanning.  Their high cost makes them prohibitive for small business use. Film scanners and combination scanners cost more than flatbed scanners.  Some flatbed scanners have film-scanning add-ons; however, these do not work very well if the media has to be laid on glass.  Glassless scanning is a must for film scanners.  High-end film scanners will usually do a better job than any combination scanner.

Some scanners can be bought for as little as $50.  They are modern price miracles.   However, do not expect a cheap or low-end scanner to perform as well as a medium- or high-end scanner.  More expensive scanners are faster and provide better color separation, pattern discrimination, and image resolution.  Here are some of the major scanner manufacturers:

1. Nikon Excellent film scanners.
2. Agfa Excellent flatbed and combination scanners.
3. UMAX A good buy for the money.
4. Microtek Good flatbed, film, and combination scanners.
5. HP Flatbed scanners.
6. EUROCORE   Expensive drum scanners.
7. MUSTEK One of the oldest scanner companies.
8. Minolta  Film scanners.
9. Linotype-Hell (Heidelberg) Flatbed and drum scanners, medium- to high-end.

Now file modes and file types will be discussed.  Four prevalent file modes for image processing are:

1. Grayscale: This mode is used for scanning black and white photos or for creative effects.  It usually encompasses 256 shades of gray, though some file formats (such as .TIF) can handle 1024 shades of gray.
2. Indexed Color: This mode indexes the colors in an image to an established 256-color range.  One use is to support transparent .gif format files.
3. RGB: This mode means red-green-blue color.  This is the most common mode for image processing.  Images displayed on a monitor are usually in RGB mode.  This mode is supported by a wide range of file types.
4. CMYK: This mode name stands for cyan-magenta-yellow-black.  CMY is the complement of RGB.  Photographic enlargers use CMY filters to adjust color balance.   The CMYK mode is often referred to as the pre-press mode, because images intended for printing are in this mode, which is why some people only process images in CMYK mode; however, there is no unique conversion between RGB and CMYK modes.  All color printers convert RGB images to CMYK images before printing them. Printers have what is called an ink limit, usually a value 270 to 290, which is prevents too much ink being used. The black (K portion) is selectively increased to cut down on the total amount of ink.

There are a multitude of raster and vector file formats.  Scanners produce raster files as a series of dots across and down an image.  Vector files are composed of polygons.  Conversion from raster to vector is difficult, but vector images can be resized and manipulated more easily.  Conversion from vector to raster is not difficult, though color smearing can occur.  Raster images must be scanned at the largest size desired, or they can develop continuity problems.  Some raster image types, specifically those that can be displayed on websites, will be discussed next:

1. .gif file: This is often referred as the Compuserve file type.  It is a lossless type, in that no information is lost from its (copyrighted) compression scheme.   .gif files can have transparent backgrounds, and they can be animated; however, they only support a maximum of 256 colors. 
2. .jpg file: This is a modified windows .BMP that uses lossy compression to reduce its storage size.  Two main types can be seen on the web: standard, which appears from top to bottom, and progressive, which appears all at once as a hazy image which gradually gets clearer.  There are a multitude of of ways for creating .jpg files, for deciding which image elements are more important than others.  Compression ratios as large as 100 to 1 are possible; however, not all image processors can handle such images.  Since .jpg compression is lossy, information is thrown away to reduce the size the file.  Agressive compression of given portion of a picture may cause banding, where there is no longer sufficient information to create gradual blending of colors.
3. .PNG file: This is a new file type that is still being developed.  It does not employ a copyrighted compression scheme, and it is intended as a standard to overcome the limitations of .gif and .jpg files.  Currently, this standard does not have a defined animation format.  For more information, see PNG - Definition and Links.
4. .TIF file: This is not a web file format.  This is a convenient format for storing scanned text to be processed by an OCR program.

See Introduction to Image Processing Next

Last modified  Friday, November 01, 2002