Compressing Color Document Images

The buzz started at AIIM '99 in Atlanta. That was when Kodak introduced a production (85 page per minute) color document scanner priced under $30,000. Since then, industry speculation has run rampant as to when color will replace black-and-white as the preferred method of capturing document images. Some estimate that in as little as two years sales of color production scanners will surpass sales of bi-tonal scanners.

There are several issues that need to be resolved, however, before color document imaging can compete with black-and-white. One involves the acceptance of a standard (or de facto standard) format for color document images. TIFF Group IV has long been the standard format for bi-tonal images. However, color TIFF files are too large to be practical in most document imaging applications.

Maintaining Textual Integrity
Kodak has embedded JPEG output into its early production color scanners. JPEG files are considerably smaller than TIFFs; however, this smaller file size comes with a cost. It is achieved through "lossy" compression, which means that some data is lost during compression. When the image is viewed, the lost data is reconstructed based on the patterns of the remaining information. Unfortunately, textual information is often damaged in lossy compression.

As an alternative to JPEG, several compression vendors have introduced color image formats that separate images' text and graphics before compressing them. To keep file sizes small, the graphics can still be compressed using lossy techniques. To maintain textual integrity, these document-imaging-friendly formats employ specialized text compression.

Microsoft Office Adopts TIFF-FX
"We have introduced a whole new paradigm for color compression," says Sebastian Gard, director of product marketing for paper management products at ScanSoft (Peabody, MA). "This paradigm is ideal for storing documents that have a mix of pictures and text, and feature rich content or a complicated layout."

ScanSoft is currently developing software to create TIFF-FX color document images. TIFF-FX is an open standard that ScanSoft and Xerox helped develop. ScanSoft has already licensed its TIFF-FX software to Microsoft for inclusion in a future version of Office.

"TIFF-FX is peer-reviewed technology," stresses Gard. "It has been approved by a standards body and adopted by Microsoft. It has also been accepted by the fax industry for color faxes."

Wireless Apps Also Demand Compression
Gard adds that the proliferation of color production scanners will drive the market for color document imaging applications. "In addition, the proliferation of wireless applications will drive color image compression software sales," he says. "In wireless applications, because of bandwidth constraints, it is imperative that files being retrieved are small."

While ScanSoft has the advantage of Microsoft as a partner, LizardTech (Seattle) has the advantage of a product that has been on the market for a few years. At AIIM 2000, LizardTech announced that it had purchased the rights to AT&T Labs' DjVu color document image compression software, which was first introduced at AIIM '99.

Better Scanners Give VARs More Options
"Because color document image compression is such a new application, it is important for VARs to work with a vendor that offers a support network," says Casey McPartland, director of global services for LizardTech. "The software can be customized to stress certain aspects. In one application, character recognition might be more important; in another, a certain color might need to be emphasized. We can help VARs leverage their knowledge of their customers' workflows into successful compression applications."

Joe Tradii, LizardTech's product manager for DjVu, adds that it is important for VARs to install quality color scanners in their color applications. "Regarding file format and DPI (dot per inch) output, it is better to have more options than less," stresses Tradii. "VARs need to convince their customers that, with scanners, you get what you pay for. A scanner should be thought of as an investment in business, not as a commodity."