National Statistics Office

---

National Statistics Office implements document imaging for Census 2000 with Fujitsu Philippines

The National Statistics Office signed up recently with Fujitsu Philippines, Inc., one of the leading systems integrators in the country today, to implement document imaging technology for capturing more than 15 million questionnaires for its Census 2000 program.

The Census 2000 program, which is currently implemented nationwide, is designed to take an inventory of the total population in the Philippines and to collect information about their characteristics.

The NSO Data Capture System calls for the use of imaging and Intelligent Character Recognition (ICR) technologies, consisting of scanners, PC's (workstations and servers), laser printers, data storage devices, and document imaging software from FPI. For the processing of the Census 2000 questionnaires, the NSO veers away from the traditional keyboard data entry method to use document imaging.

ICR Technology

"The decision to use ICR technology in our Census 2000 is a bold move on the part of NSO," says NSO Administrator Tomas Africa. "With this move, the Philippines joins the ranks of the few countries in the world utilizing this new technology. While we cannot match the sophisticated systems of the advanced countries like the US and Australia, our adoption of ICR nevertheless puts us a notch closer to these technology-oriented countries."

He adds that in Asia, he believes the Philippines is one of the first countries, if not the first, to use ICR for the census, Indonesia being ahead of the country to use ICR for the census. Other countries, encouraged by the country's move to use ICR technologies, may also adopt this solution in their respective censuses either late this year or next year.

Risks in new technologies

The NSO administrator admits that there will always be risks in using new technologies such as ICR. "But we at NSO also did our homework before we made the bold decision to use the ICR technology. If we look back at how we managed past censuses, we see a history that is marked with bold moves with respect to new technologies," he says.

In 1990, for instance, NSO used standalone computers in a decentralized set-up to process the census. Other countries, including the more advanced ones, thought that this was impossible. "But we proved them wrong. In the end, the Philippines got the distinction of being one of the first countries with population of 60 million to process its census using entirely microcomputers."

In 1995, NSO implemented some technological innovations by introducing networking during the data entry phase and in the transmission of data. NSO Administrator Africa claims, "There had been problems each time but we managed to overcome them. As we embrace yet another technology, we are confident that we can again manage the risks and make the project successful, especially because we have a global company such as Fujitsu being our partner in this census round."

"We've only started our partnership with FPI. We hope that we will be able to bank on and build on this partnership in the weeks and months to come, until we release the final population counts," he adds.

Fujitsu partnership

NSO Director, Valentino Abuan, said the bidding process for the NSO DCS project started in August last year with nine bidders, including FPI, submitting their technology proposals. After the pre-qualification bid, five bidders were qualified and FPI was one of the chosen five.

"When we opened the bidding for NSO Data Capture System project and after careful evaluation, FPI offers a better proposal with regard to the overall total cost of ownership. Cost was just one of the factors," Abuan added.

Felipe R. Manalang, EVP and GM of FPI, for his part said: "The real asset of FPI is its IT workers' practical skills in the implementation, operations, and maintenance of new technologies. FPI believes in using its knowledge, rolling up its sleeves, and doing the job as technology practitioners and implementers to meet the business goals of its clientele."

The NSO Data Capture System - How does it work?

Data capture, or the process of converting the entries in the Census 2000 forms into machine-readable format, starts with the scanning of every page of the accomplished census forms, which have been earlier batched at the Provincial Offices, to create an image for every page.

The resulting images are passed on to a recognition program which handles the automatic conversion of optical marks (OMR), machine-printed information (OCR) and hand-written information (ICR) into ASCII text.

The automatic conversion of page images to ASCII text is called the data recognition stage. Those pages or fields, which cannot automatically be converted to ASCII, are resolved manually or re-typed by the operators in what is known as the data resolution stage. Once the fields have been resolved, the ASCII texts are saved into ASCII files for further processing of the data.

The Data Capture System (DCS) consists of scanners, PCs (workstations and servers), laser printers, data storage devices, and the Mid Volume Capture Scanner (MVCS) and Eyes and Hands for Forms (EHF) software. Each capture workstation has MVCS Capture Software installed on the workstation. Kodak Digital Science Document Scanners 3510D and 3590C are used in scanning the forms and are attached to the workstations with the MVCS Capture Software. MVCS is involved solely in capturing the forms and data electronically.

Interpretation, verification and transfer of the images and data are the functions of EHF. The EHF Software is installed on the workstations. The software is tasked to do the interpretation, verification, and transfer of images and data. For networked installations, EHF makes use of Velocis, a type of database engine, which is installed on the Image Server.

Images and data captured by the MVCS Capture Software are saved on the Image Server. The Image Server also handles the load balancing of the processes or modules running in each workstation. The software running on the server manages the storage of the images and data through the process. The Image Server acts as intermediary when a workstation needs to obtain the images to be processed and the repository for the data to be verified.

Each EHF software then imports these images and data to be interpreted, verified, and then later transferred by the appropriate EHF production modules to the specified system. Laser printer and data storage devices like MO Drive and CD Writers may be attached to any one of these workstations.

NSO has four Data Capture Centers nationwide that include Manila, Pampanga, Cebu, and Davao.

---