Previous | Table of Contents | Next |
Gilbert Held
Careful planning is required so that the burgeoning use of image-based applications results in enhanced user productivity rather than decreased network performance. The several techniques reviewed help tailor an images characteristics to the intended application to optimize network bandwidth and server storage availability.
Advances in monitor display technology coupled with the development of multimedia software and increased capacity of disk drives are just some of the several factors contributing to a rapid increase in the use of images in client/server environments. Today it is common to find pictures of employees in a personnel database, images of houses and interiors in a real estate database, and the results of CAT and MRI scans in a hospital patient database.
Although the adage one picture is worth a thousand words contributes to the increased use of images, the storage and transmission of images can rapidly diminish network resources. Using images in client/server environments has to be carefully planned so that the resultant application is effective both in terms of cost and of operations. If it is not, the potential adverse effect to network performance will decrease the productivity of all network users those using image-based applications as well as those using other client/server applications. Inappropriate use of images can also result in inefficient use of disk storage, necessitating costly disk drive upgrades or the acquisition of a hierarchical storage management system whose use introduces delays as migrated files are moved back to disk storage on the receipt of a user access request.
Appropriate use of images is therefore vital to enhancing user productivity and the effective use of network bandwidth and storage.
Two basic types of images are used in computer applications: raster and vector. A raster image consists of a grid of equally sized pieces referred to as pixels and records a color element for each pixel. Raster images include those taken with a camera and scanned photographs. In comparison, vector images are images resulting from a collection of geometric shapes that are combined to form an image and are recorded as mathematical formulas. Computer-aided design (CAD) drawings represent an example of a vector image.
Vector data cannot reproduce photo-realistic images. As a result, most computer-based applications require pictures of persons, places, or things using raster-based images. This discussion of techniques for using images effectively in a client/server environment is therefore limited to raster-based images.
A color is associated with each pixel in a raster-based image. That color can vary from a simple black or white denotation to the assignment of one color to each pixel from a palette of more than 16 million colors.
The assignment of a color to a pixel is based on the use of one or more bits per pixel to denote the color of each pixel, a technique referred to as the color depth of the pixel. For example, a black and white raster image would use one bit per pixel to denote the color of each pixel, with each pixel set to 1 to denote black and 0 to denote white. For a 16-level gray scale raster image, each pixel would require 4 bits to indicate the pixels gray level (24 = 16). Exhibit 1 indicates the correspondence between the number of bits per pixel (i.e., color depth) and the maximum number of colors that can be assigned to a pixel. Note that the use of 24 bits per pixel is referred to as true color and represents the maximum number of colors the human eye can distinguish when viewing a raster image.
The amount of storage required for an image depends on its size, resolution, and color depth. The size of an image references its vertical and horizontal size typically expressed in inches or millimeters. The resolution of an image references the number of pixels per inch or per millimeter, and the color depth represents the number of bits per pixel required to define the color of each pixel in the image. Thus, the total amount of data storage can be expressed in bytes as follows:
Exhibit 1. Correspondence between Bits per Pixel and Maximum Number of Colors | |
---|---|
Bits Per Pixel (Color Depth) | Maximum Number of Colors |
1 | 2 |
2 | 4 |
8 | 16 |
16 | 32,768 or 65,536 (depends on format) |
24 | 16,777,216 |
Exhibit 2. Data Storage Requirements for 3.5 × 5-Inch Photograph | |
---|---|
Data Storage (Bytes) | Color Depth |
196,875 | 1 bit (black and white) |
393,750 | 2 bits (4 colors/4-level gray scale) |
787,500 | 4 bits (16 colors/15-level gray scale) |
1,575,000 | 8 bits (256 colors/256-level gray scale) |
3,150,000 | 16 bits (32,768 or 65,536 colors) |
4,725,000 | 24 bits (16,777,216 colors) |
Previous | Table of Contents | Next |