Previous Table of Contents Next


A lossless compression method is fully reversible and does not result in the loss of any details on an image. In comparison, a lossy compression method groups blocks of pixels and considers two blocks to be equivalent for compression purposes if they differ by only one or a few pixels. Thus, decompression can result in the loss of a few pixels per block of pixels. Lossy compression can significantly reduce the data storage requirements of images. Although this compression method would not be suitable for storing some images, such as a chest X-ray where every pixel is important, it is highly effective for storing pictures of personnel, homes, and other objects for which the loss of a few pixels does not significantly alter the image’s usability.

Exhibit 3. Comparison of Data Storage Required for Four Different File Formats
File Format Data Storage

TIF no compression 2.4M bytes
GIF LZW compression 1.8M bytes
JPG least compression 1.365M bytes
JPG moderate compression 163K bytes

To illustrate the potential differences in data storage obtained through the use of different file formats, the same photograph was scanned several times and stored each time with a different format. The first scan was accomplished specifying the TIF file format without compression. The resulting file required 2.4M bytes of data storage. Next, the same image was stored using the CompuServe GIF file format, which includes built-in LZW compression. The resulting data storage required for the photograph was reduced to 1.8M bytes. Although a reduction of 600K bytes is significant, LZW compression is fully reversible and does not represent the best method of compression for storing many images.

To illustrate the potential of data storage reduction that can be obtained through the use of lossy compression, the photograph was rescanned two additional times. The first time the image was scanned using the JPG file format specifying least compression, which results in two blocks of pixels being considered as equal if they only differ by one pixel. The resulting scan required 1.365M bytes and yielded a reduction of 435K bytes in storage from the GIF file format. Next, the image was scanned and saved using the JPG file format specifying moderate compression, where two blocks of pixels are considered to be the same even if they differ by up to four pixels per block. As a result of using a moderate level of lossy compression, the amount of data storage was reduced to 163K bytes — a reduction of more than 2M bytes in the storage of the image using a noncompressed file format. Exhibit 3 provides a comparison of the data storage required for storing the same photograph using four different file formats.

If each file was stored on a server on the previously described Ethernet LAN, it would take 1.365M bytes x 8/250K bps or approximately 44 seconds to retrieve the image stored using a JPG file format and least compression. In contrast, the retrieval of the image stored using a JPG file format and moderate compression would require 163K bytes x 8/250K bps or approximately 5.2 seconds. Thus, the use of a moderate level of lossy compression can significantly reduce client retrieval time as well as server data storage requirements.

Cropping

Another technique that provides an effective mechanism for reducing image storage and transmission requirements is cropping. Cropping involves eliminating portions of an image that are not applicable to the application. For example, if images are required for an employee database, a large portion of the background of the scanned photograph can be cut out.

File and Image-Type Conversions

Two additional techniques for enhancing the use of images in a client/server environment are file and image-type conversions.

File Conversion. Many times software provided with a scanner or digital camera limits the user to storing images with one file format. Thus, the use of an image management program that provides a file conversion capability may provide the ability to significantly reduce the data storage for an image. By converting from one file format to another as well as specifying the use of an appropriate image compression method for the application, users can significantly reduce the storage requirements for an image.

Image-Type Conversion. The second conversion method involves a change in the image type. Here image type is used to denote the use of color depth with an image. For example, assume a photograph is scanned using 24-bit true color. Three bytes would then be used to store color information for each pixel. If the photograph is being used in a personnel file, the image could probably be converted to 16 color requiring 4 bits per pixel or 256 color requiring 8 bits per pixel to store color detail information. Thus, converting a one-inch 300 dpi image requiring 90,000 pixels from true color to a 256-color image type would reduce its data storage requirement to 180,000 bytes.

CONCLUSION

Although the use of images in applications has the potential to enhance user productivity, it also can adversely affect network bandwidth and server storage availability. Effectively using images in network-based applications requires the careful consideration of the use of different image file formats, compression methods, cropping, and image type. By carefully considering each of these image-related features, the characteristics of the image can be tailored to the application. Doing so can significantly reduce the amount of storage required for the image as well as the transmission time from servers to client workstations.


Previous Table of Contents Next

Copyright © CRC Press LLC