InputAccel:
By Kimra Hawley, Vice President, General Manager
Software Tools & Input Subsystems Division, Cornerstone Imaging, Inc.
Document capture, the conversion of paper documents into digital images for data extraction, is a complicated, labor-consuming process essential to the operation of an imaging system. Although hardware and software are purchased once, labor costs continue for the life of a system. As a result, some companies have discovered that over a period of years their document capture labor costs exceed their initial investment in imaging system equipment.
Solutions Must Accommodate Unique and Evolving Needs
Selecting the right document capture system can help minimize these costs and allow you to recoup your imaging system investment over a shorter period of time. But to do so, the system must be: 1) customizable 2) open 3) manageable 4) upgradable. In short, it must offer you choices, because only such an adaptable system will be able to accommodate the specific needs of your imaging application.
1. Customizable. Effective document capture solutions must be customizable in the following ways:
* Technology. Different applications require very different technologies. For example, some capture systems need OCR, handprint recognition or forms extraction, while others don't. Some need lots of image enhancement, others don't need any. You need the flexibility to choose the right set of technologies for your specific capture application.
On-going document capture operational costs can exceed initial capital investment in imaging system hardware and software.
* Products. Different products, even within the same technology category, perform very differently. For example, some OCR engines work best on fixed pitch machine print, others on variable pitch. Some work well with columnar text; others don't at all. Some image enhancement products do a good job cleaning up images and reducing file sizes, but have a negative effect on OCR accuracy. And so on. The right product for the job depends on your document stream characteristics and overall system goals -- there's no one best product in any technology category.
* User Interfaces. The user's view into the system must be customized to maximize operator productivity and achieve system goals.
2. Open. To be truly customizable, the document capture system must be open. That is, it must provide the freedom to choose the right technologies and the right products for the specific application, even if these products are provided by several different vendors.
3. Manageable. Selecting the right technologies, the right products, and customizing user interfaces isn't sufficient to yield a production document capture solution. Advanced system management and reporting features are needed to transform a collection of technologies and products into an effective, manageable system. Features should include:
* Customizable capture process. Different applications require that pages flow through the system in different ways. For example, some pages may go through OCR processing, others need specialized image enhancement and still others require rescanning. Productive capture systems require an intelligent work queue manager to manage and control routing of pages throughout the system.
* Process flexibility. Many document capture systems require all documents to move in clumps through the system -- for example, all pages must complete scanning before image enhancement can begin. They then must complete the enhancement process before key data entry or OCR starts, etc. Human and machine resources are wasted when left waiting for the next batch of work to arrive. An effective capture system allows pages to move on to the next step as soon as a previous work step is completed, without waiting for the rest of the batch.
* Reporting tools. Reporting tools provide the machine and operator information you need to optimize, tune and control your system. Operational information, such as the percentage of pages that need to be rescanned, number of flagged OCR errors that are real errors, status of a given page at any point in time, number of forms each data entry operator completes in an hour, etc., is critical.
4. Upgradable. Your document capture needs will most likely change over time. You need the flexibility to easily upgrade the volume and functionality of your system.
InputAccel document capture system
example.
* Volume. Most document capture systems begin with a pilot system which converts relatively low volumes of documents. When the project moves to deployment stage, the capture system needs to easily scale up in volume.
* Functionality. As your needs change or as new technology becomes available, the ability to add new capabilities to the system is important.
Document capture solutions that meet these four criteria will help you manage the costs associated with document capture and recoup your imaging system investment over a shorter period of time.
InputAccel's Offers Flexible Building Block Approach
InputAccel uses a building block design that results in document capture subsystems with the customization of a one-of-a-kind system without the high development costs. InputAccel offers the simplicity and ease-of-use of a shrink-wrapped solution without the performance limitations. It also offers the ability of toolkits to integrate diverse products without a sacrifice in system management capabilities. In short, InputAccel puts users in control of their own destiny because its building block design makes it easy to:
* hook together the various technologies and products needed to build a document capture system.
* customize the application to fit users' specific needs.
* effectively manage and control the resulting system.
InputAccel offers the simplicity and ease-of-use of a shrink-wrapped solution.
InputAccel is an open system, meaning it can use technologies provided by Cornerstone, by other vendors, or of an integrator's own creation. Solution providers are able to choose the best available technology for virtually any image or data capture task. In fact, the following companies have announced plans to provide InputAccel-compatible products:
OCR/ICR Products:
AEG
Caere Corporation
Electrocom Imaging
Mitek
Nestor Systems
Prime Recognition, Inc.
Image Processing Products:
Dunord Technologies
Picture Elements, Inc.
Seaport Imaging
Sequoia Data Corporation
T.i.S.
Forms, Data Entry Products:
Comspec Corporation
Datacap, Inc.
DMAC
GTESS Corporation
Recognition Research, Inc.
TeraForm Corporation
TextWare Corporation
Wheb Systems
Special Modules:
Adobe Systems
Diamond Head Software, Inc.
ICONOVEX Corporation
IMR, Inc.
Applications:
Imagery Software
Lotus Development Corporation
Recognition International, Inc.
ViewStar Corporation
Zylab
Image Architects
InputAccel's open design makes it easier to integrate appropriate technologies into a custom document capture solution that can be easily upgraded to accommodate changing needs or new technologies. With InputAccel, many organizations will find it economically feasible to install sophisticated document capture systems to improve imaging worker and system productivity.
As shown in the figure, an InputAccel system is composed of :
* InputAccel/Server--a framework for controlling and administering the document capture process. It is the backbone that connects all the parts of InputAccel. It manages the flow of work through the capture process, transparently balances work across available resources, and serves as a repository for system status and reporting information.
* InputAccel-Compatible Modules --functional modules (or clients) that manage specific document capture tasks are supplied by both Cornerstone and a large set of InputAccel partners. Integrators and application developers are able to choose the best available capture technology for virtually any task, or can even develop their own InputAccel-compatible modules. InputAccel-compatible modules can be integrated into the system without programming, and only the modules required for the specific document capture application need to be installed.
Currently, in addition to modules provided by other companies, available modules include:
* InputAccel/Scan
* InputAccel/Image Enhancement
* InputAccel/Indexing and Quality Assurance
* InputAccel/Re-scan
* InputAccel/OCR
* InputAccel/OCR Edit
* InputAccel/Image Export
* InputAccel/Index Export
InputAccel Automates Test Scoring at CTB-McGraw Hill
As the nation's leading elementary and secondary school test publisher, CTB-McGraw Hill provides a variety of assessment, scoring, and reporting services to schools and state education agencies. Based on the ease of system integration and the open architecture of InputAccel, the company decided to adopt document imaging as a way to expedite the scoring of tests.
One of the intended uses was to automate the routing of tests for scoring. Previously each scorer was responsible for grading a specific set of questions on the test and then manually forwarding the test to the next scorer. With InputAccel, test pages are scanned and then automatically indexed using OCR and barcode recognition to identify the test number and student's name. Indexed tests are then exported to FileNet workflow software which electronically routes them to the appropriate test scorers. Different sections of a single test can be scored simultaneously, eliminating the need for sequential scoring of test sections.
Scanning volume for the InputAccel system is expected to reach upwards of 10,000 pages a day, with 300,000 pages a day required during peak test times. According to Anagh Chatterjee, CTB's director of technology, "Use of this and other such enabling technologies will facilitate maintaining and enhancing our position as a full-service test publisher offering both testing products and scoring services."
InputAccel Used to Consolidate Patient Records at Stanford Hospital
Stanford University Hospital installed a pilot InputAccel system to capture upwards of 1,000 pages a day of records for tests and services performed in the hospital's labor and delivery unit. Captured images were exported to FileNet workflow software. In the current second phase of the pilot project, the system is being distributed throughout the hospital so that 10,000 to 20,000 pages a day can be scanned in 15 to 20 different locations. All images are exported to the same document management system.
Bill Hester, Stanford's Technical Project Leader for Business Process Re-engineering, said, "InputAccel was chosen because it allowed us to capture document images throughout our institution, at various physical locations, while maintaining quality control over the images." Hester also noted, "The product's scalability and support from a stable and well-known company provide enormous benefits to the end customer as well as the systems integrator."
InputAccel Provides Backbone for a Backfile Conversion Service
In September 1995, CADPAC, a 13-year-old service bureau headquartered in Cincinnati, OH, installed an InputAccel document capture system as the basis for a new scanning service bureau. Following installation and training by Cornerstone, CADPAC used the InputAccel system to begin converting 500,000 pages for the Customer Service area with the Automotive Division of Rockwell International.
Steve Naghshineh, CADPAC's Manager of Imaging Solutions, said that his selection of InputAccel for the new scanning service was based on the system's flexibility. "Every customer wants its images produced in a different way. There is no standard in this area. In the imaging world, change is the constant. Routing, scanning, and workflow tasks often will change daily. InputAccel was the only system I saw that provided us with enough flexibility to make changes on the fly." He also liked the fact that, "You don't have to be a computer guru to run this product."
The Rockwell project site was set up in two days and conversion began immediately. According to Naghshineh, "We went in at 7:00 a.m. and began scanning two hours later." He promptly discovered that his requirement for system flexibility was well justified. The Rockwell documents were both letter and legal size and of variable quality. CADPAC workers soon found that the InputAccel/Scan module, as initially configured, required more mouse movements and operator intervention than they desired. In a matter of hours, the required changes were made and scanning resumed at a rate of 8,000 to 10,000 pages per day.
Nashshineh noted another InputAccel benefit: "We can ramp up according to our needs, not the vendor's. Cornerstone's InputAccel pricing reflects the number of pages being scanned daily." As of December 1995, CADPAC had new conversion projects scheduled for its InputAccel system through July 1996. *
Kimra Hawley, vice president and general manager of Software Tools and Input Subsystems at Cornerstone Imaging, Inc., has over 14 years of experience managing computer systems development and marketing teams. In here current role, she leads the Cornerstone division that develops InputAccel and the Pixel Translations PixTools image product line.
IW Special Supplement, March 1996
|
![[LiveLink]](/file/36067/SIGCAT 97 - Conference on Convergence.ISO/Docs/webcd.gif "[LiveLink]"){kind=small}