Increase Image Quality While Decreasing Labor

The value proposition for scanning and retrieval of documents is pretty much a given - as long as the documents lend themselves to being scanned. Aging documents, faint type, color annotations, and other factors can make the process of scanning some documents very labor-intensive and require adjustment of scanner settings and multiple rescans to get an acceptable image. One alternative is to scan in color to capture all of the details, but that presents its own set of challenges. Most OCR/ICR (optical character recognition/intelligent character recognition) engines can't read color images, so even performing zonal OCR for automating indexing could require a conversion to bi-tonal or color drop-out at some point. Color or grayscale images, even of black-and-white documents, are larger files than bi-tonal ones. "A bi-tonal image might be just 20 to 50 KB," says Al Hawkins, president of document service bureau OptiScan (Phoenix). "But a full grayscale image could be as large as a megabyte." Even with falling storage prices, some organizations may find the volume cumbersome and have concerns about the demands of transporting images across a network or via e-mail.

It was this dilemma that prompted OptiScan to create Scantastic, a software package that manages client documents from the capture process until they are exported to CD. Scantastic controls scanner settings and quality control procedures, merges indexing information into a database, and exports jobs in various formats with volumes sized for CD distribution. In March of 2002, Scantastic underwent its toughest test when trying to fulfill the requirements of a contract with Honeywell. The aeronautics giant retained OptiScan to convert decades of existing engineering and technical documents to digital images.

Dual Streaming Provides Image Quality Choices
Honeywell's documents presented a number of challenges. The paper was aging and faded in many cases. Some were on blue-and-white graph paper. Designs had been modified over the years, items written in pencil had been erased, and handwritten notes had been added. Because of the potential importance of this data, Honeywell set forth an uncompromising quality standard: every element of every document must be completely legible. OptiScan needed to find a way to meet that requirement without extensive labor costs. "What we had feared was that we would be in a position of marking massive quantities of documents for rescanning," recalls Hawkins. "The dual streaming capability of Panasonic's [Seacaucus, NJ] KV-SS905C scanner offers us a better way to get good quality images without having to sit down and pull pages out. This cost savings and efficiency allowed us to bid at a competitive price."

The 174 ipm (images per minute) KV-SS905C simultaneously creates bi-tonal and color images and can send both to the software application. After the documents are scanned, the images are visually reviewed for quality assurance. The verifier looks first at a black-and-white image, which would create the smallest file. If it is completely legible, the image is sent to the export process. If not, the grayscale image is selected and reviewed. When no other format will capture all of the details, the color image is saved. "When scanned in black and white, graph paper with blue squares, for example, left lines through the document, making it impossible to read," says Hawkins. "Using grayscale conversion, those lines fade into the background. The documents are legible, and the file size is still smaller than a color image." Without dual streaming, the scanner would have to do all bi-tonal images or all color images in separate batches. Manually separating the pages to run them in different batches requires additional labor and opens the possibility for errors when stacks of documents must be reassembled.

After all of the images have gone through the quality check, Scantastic software puts them in their original order and converts them all to PDF (portable document format). The images are used by several divisions at Honeywell, including manufacturing and purchasing. "Since they have created so many parts and many of them are on airplanes that are still flying, that data has to be available to a wide range of users," says Hawkins. "The plan is never to go back to paper." Honeywell uses a product called PixEdit from Norwegian software developer Techsoft to view, edit, and annotate color or bi-tonal documents. "They can erase or edit and output a new version," comments Hawkins. "They no longer have to print and rescan documents which could compound any quality issues."