Capture family histories
AncestryDPS, a historical records specialist, has captured more than 5 billion images related to family histories since 1996. The company, based in Provo, Utah, works with governments worldwide to convert into digital format such documents as census, birth/death, military and immigration records.
The firm creates indexes for the 100,000 images it captures each day. For forms, it uses specific fields, such as name and date of birth. For non-formatted documents, such as pages of a local history book or an obituary in the newspaper, it creates a full-text index.
Historically, AncestryDPS relied primarily on a custom capture system that consisted of optical character recognition and intelligent character recognition. But the company knew it needed a capture solution that could better handle its growing volume of documents. It also wanted a solution that could leverage the technology the company already had.
"We sought an application with robust workflow that could help us track and process a higher volume of data and documents," says Shawn Reid, AncestryDPS' director of development. "We were also looking to increase the amount of automated capture we could perform on documents with machine print.
Because AncestryDPS deals with documents from such a wide variety of sources, the ability to efficiently customize its capture application was paramount. "Most businesses deal with the same types of documents over and over again," says Brandon Groom, CTO of value-added reseller Doxtek of Orem, Utah. "Every time AncestryDPS gets a new set of books, it is dealing with a new type of document."
"We process dozens of document types each year, and when we finish with one, we may never see it again," Reid adds.
The majority of the documents are captured from microfilm images, and some of the originals are more than 200 years old. As a result, AncestryDPS incorporates some unique image processing algorithms.
The company eventually decided to purchase the Kofax Ascent platform and Kofax's INDICIUS technology. Ascent and INDICIUS have been deployed to handle automatic information capture from forms with machine-printed data, which account for about five percent to 10 percent of AncestryDPS' overall document volume.
"On those documents, we have reduced our capture costs as much as 10 times, because of reduced keying requirements," says Reid.
Reid adds, "INDICIUS has improved the efficiencies of our keying operations through features like tick lists and automated validation. For example, INDICIUS enables us to run automated cross-checks, such as verifying that someone listed as a son on one part of a census form isn't listed somewhere else as a female."
AncestryDPS has also deployed the Ascent Collection Internet Server (ACIS) to manage its interaction with its keying subcontractors. Furthermore, it plans to leverage Ascent and INDICIUS to expand its business into more general information capture, according to Kofax.