July 1998

Schools Give Imaging An "A"

Imaging isn't just for scoring tests anymore. It's being used to provide classes over the Internet, to create digital libraries and to help students get jobs in the real world.

Remember taking the SAT and having to fill in all those rectangles? Schools have been scoring tests using scanners and optical mark recognition for years. Now they're branching out into other uses of document technology. Here's a look at three leading-edge educational applications: developing digital libraries, training students for multimedia careers and teaching courses via the Internet.

1850s Journals
Brought to Life

If there's anything you want to know about what life in America was like during the period of 1850-1877, go to www.umdl.umich.edu/moa and you should be able to find it. This is where the University of Michigan has posted an impressive digital library of books and journals that document American life leading up to and following the Civil War. The collection of 5,000 volumes includes diaries, first-person travel accounts and popular magazines.

The Web resource provides an alternative perspective that you won't find in military histories or political tomes. The purpose is to show what it was like to be an American at that time. The books and journals were already part of the university's physical library.

"We're one of the largest research libraries in the world," says John Price-Wilkin, head of digital library production service. "We've got about seven million volumes, including some that are extremely brittle. Most of them haven't circulated since the 1950s, and now they're so actively used it's amazing. We've had about 50,000 pages viewed each month over the last three months for books that were [formerly] in cold storage."

A service bureau scanned the books and magazines for this project and created 600 dpi bitonal TIFFs. Price-Wilkin declines to name the bureau because the university found them unresponsive. They were also disappointed with the original optical character recognition (OCR) program they used to pull the text off the images.

But there's a happy ending to this story. The richness of the 600 dpi images let the university run them through a second pass under more effective software -- this time, PrimeOCR from Prime Recognition (San Carlos, CA 415-637-8382). This image cleanup and OCR software deskews the images, enhances them with despeckle, line removal and image rotation, and recognizes them using four OCR engines and a voting engine.

"We just finished a second round of OCR using Prime Recognition software and it's been immensely better than the results we had the first time," says Price-Wilkin.

The OCR accuracy rate averages 99%, which is helpful to the students. "Many people want to see the text and not the images. There are about seven volumes in there that are fully corrected and coded. They look like HTML pages and provide links to the images. That's generally the way people have preferred to do things. It takes less time to transfer a chapter of text than it does a single-image page. There are a lot of users who are vision impaired, and the text is more readable for them."

The university rescanned any images that were unreadable using a Xerox 620 scanner. The images were stored as TIFF files, but because most Web browsers don't display TIFFs, a former university employee wrote a tif2gif utility that converts the images to GIF format on the fly as people retrieve files over the Internet. The conversion happens so quickly that the site visitor can't tell it's happening.

The files are stored on two MTI RAID systems. In a digital library of this size, a good search engine is important. The University of Michigan has been using OpenText's search engine since 1989. (The search engine isn't sold separately any more; it's now part of a document management system.)

When looking at search engines, "We look for extremely fast retrieval from large bodies of information," says Price-Wilkin. "Our community is not so interested in fuzzy logic. We're interested in high precision--the exact word. I want to see when people have spelled Kansas with a "z." I want to see how the language has changed. Fuzzy logic gets in the way of doing that. Another thing that's big in our community is phrase searching. You could do a search on 'to be or not to be' in our project and get a very speedy response."

The scanning for this project cost about 12.5 cents per page. The OCR -- including equipment and staff -- cost 10 to 15 cents per page. Building the system, including people and equipment, cost roughly $10,000. The RAID systems came to about $45,000.

The University of Michigan is talking to other institutions about helping them with their digital libraries.

Starting Students on
Technology Careers

Keith Rutledge, a multimedia teacher at Travis High School in Austin, TX, wanted to invent a program that would help inner-city kids find jobs. He interviewed business people at several of the hundreds of multimedia companies in the Austin area. He found out what their needs were, what their processes were and what software they used. Then he designed a curriculum around what he discovered.

The course teaches ninth through twelfth graders the whole process of creating multimedia, from brainstorming to storyboarding to handling the production work, beta testing and CD mastering. The classroom is mostly an Apple (Cupertino, CA 408-996-1010) shop. They have a mix of older Macintoshes like 575s and PowerMacs. The two newest machines are G3s.

"I want to have a lab completely full of G3s eventually," Rutledge says. "They're screamers."

The 28 students in the class have been sharing 16 Macs. In theory, that's good for promoting teamwork. In reality, it's a challenge.

The students capture images using an Apple QuickTake digital camera and a Kodak (Rochester, NY 716-724-4000) DC50 camera. They scan photos with an Apple Color One Scanner, and they're hoping to buy a 4x5 transparency scanner.

They use Adobe (San Jose, CA 408-435-8900) PhotoShop to manipulate the images. Their rendering software is MetaCreations' (Dublin, Ireland 353-1-662-9333) Bryce 3D and Infini-D. They use Apple's QuickTime VR authoring software. When they've created a multimedia presentation, they burn it onto CDs using Adaptec's (Milpitas, CA 408-945-8600) Toast CD-R authoring software.

The equipment he's bought so far comes to about $70,000. When the budget permits, Rutledge says he plans to buy Electric Image's (Pasadena, CA 626-577-1627) Electric Image and auto-des-sys' (Columbus, OH 614-488-8838) form-Z for 3D rendering.

As part of the course, the students take on projects for nonprofit organizations. They created a kiosk presentation for Texas Parks and Wildlife. They prepared a multimedia tour for the Children's Museum in Austin. They've created interactive CDs that teach kids about history, physics and dinosaurs. A graduate student who interned with the course for a year helped students create a children's animated CD version of the Charlotte Diamond song, "I Want To Be A Dog." The students are working on a virtual reality Web site for the Institute of Texas Cultures in San Antonio.

Rutledge organizes student internships at nearby businesses. If the kids do well, the companies often hire them once they graduate.

"I've got kids making $25,000 to $30,000 right out of high school," Rutledge says, adding that at least three companies have hired students for full-time jobs.

Around the World in 30 Sites

Distance learning--the acquisition of knowledge and skills through mediated information and instruction, encompassing all technologies and other forms of learning at a distance--is the perfect place for imaging. A biology teacher can refer a class to the Web site of a zoo that posted scanned images and information about exotic animals, for example. Training materials on more than 100,000 subjects are available on the Internet.

The Benjamin Franklin Institute of Global Learning in San Diego matches English-speaking teachers with overseas job openings and helps them find online resources to make their jobs easier. One teacher who planned to teach English in the Czech Republic needed a teaching certificate. The Institute researched it and found that he could get the certificate more cheaply in the Czech Republic than he could at home.

"We tell people who go abroad about the huge library of information out there [on the Internet]," says John Hibbs, founder of the Institute. "We steer them toward the sites that we're comfortable with and that we think are appropriate for them, their mission and the country they're going to."

The teachers Hibbs works with usually offer imaging-related distance learning tools to students who live too far away to go to a class, who are disabled or who may need extra tutoring outside the classroom.

Email is one of the best distance learning tools. Students' questions and teachers' answers can be shared with the entire class.

"There are many students who don't want to raise their hand in class," Hibbs points out. "They're a little shy or they've thought of something after class that they want to ask about while they're studying at home. Email gives the professor time to answer a question more thoughtfully than he might in the classroom."

The Internet is another great tool for long-distance learning. "To be able to tell the students, 'go look at this site' is like having a giant Xerox machine that lets you instantly show copies to everyone," Hibbs says.

Entire courses can be Webcast via the Internet. A multimedia course is currently being taught to students at USC San Diego from a college in St. Petersburg, Russia, via the Internet. The arrangement came about through Global Learn Day, a 27-hour Internet broadcast hosted by Benjamin Franklin Institute and facilitated by Education News & Entertainment Network (San Diego, CA 619-503-5551) October 10 and 11. ENEN used its NetEvent service, which lets organizations broadcast live multimedia presentations over the Internet with audio, video and interactive chat between speakers and participants.

The live Internet conference started in Guam and moved through 30 countries during the day, including Brazil, Russia and the U.S. The event showcased 40,000 distance-learning courses. Anybody with a 386 PC or a Macintosh II with a 14.4 modem could participate.

Participants visited the Benjamin Franklin Web site (www.bfranklin.edu) and entered their email address to join the event. Once enrolled, they could listen in real time and browse the speaker's biography.

These are a few ways imaging technologies are used to teach students. Imaging also plays a big role behind the scenes at schools. Applications and tuition forms are scanned and processed electronically. Roommates are matched with scanned forms. And SAT tests are still graded through scanning and optical mark recognition.