Technical Overview

BSCAN Overview

BSCAN Capture was first sold in early 1994 to image capture service bureaus and as a front-end to image management and archive software. The product remains focused on high-speed capture of large volumes of documents.

BSCAN Capture systems consist of one or more personal computers, usually networked, each running a single copy of BSCAN software. A single station can perform one, several or all of the capture functions required for the application. These functions include high-speed scanning, QC, image processing, re-scanning, automatic, semi-automatic and manual indexing and releasing to another BSCAN station, formatting for archive or distribution on CD/DVD or exporting to an image management or archival system.

BSCAN's Production-Level User Interface

BSCAN provides two general interface styles: 1) a pre-defined graphical interface; and 2) a customizable interface. The pre-defined interface is typically used to perform administrative functions such as configuration, setup and monitoring while the production scanning and post-processing workers normally use an interface that is stripped of unused functions and optimized for the tasks being performed.

The pre-defined interface allows navigation between projects, applications and jobs, configuration and setup of applications, job management and scanning, QC, image processing and indexing functions. Certain classes of functions can be disabled.

The customizable interface can either enhance or completely replace the pre-defined interface. Buttons and keys, menus, dialog boxes, window and desktop layout and scripts for both control and image and index processing can be built without a special development environment. A typical configuration for scanning and image cleanup can be built from scratch in less than an hour. Unique and unusual indexing schemes can be built using BSCAN's rapid data correlation facility.

BSCAN's 'Open Architecture'

The term 'open architecture' can mean many things. For image capture systems, a product that interfaces with more than one type of scanner and outputs more than one format of image and index information can be called 'open'. BSCAN designers use the following definition.

An open system should offer as many points as is reasonably possible to modify the data being processed by the system and as many points as is reasonably possible to initiate, modify, terminate and replace elements of the system. These access points should be compatible with operating system facilities and as many independent software applications as is reasonably possible and should not duplicate existing operating system facilities except when significant added value can be realized.

Flagship Technologies

From the time of its initial release in early 1994 to today, as new image reading and processing engines have become available, selected engines have been added to the BSCAN products. In the same period, new engines have been developed by Image Access to enhance or replace existing engines. Among the automated indexing engines most often used are BSCAN's barcode and trained text readers and its pattern recognition engine. It appears at this writing that the barcode engine is the best available barcode reader for damaged and out-of-specification barcodes in the document imaging industry. It reads forward, backward, horizontal, vertical, diagonal barcodes, including barcodes that are very damaged.

The trained text reader specializes in reading difficult to read text with a high success rate and a virtually infallible confidence factor. Once the engine is trained, small and damaged dot matrix text is often read with a 99.5% success rate. If reading a six digit invoice number printed through the first part of a 2 part form with a dot matrix printer, two in one hundred may be reported as suspect with one in one hundred actually requiring correction. For documents with tight registration requirements (e.g. a small box in which the text appears), form registration can be performed at numerous points on a page to remove the distortion that can be introduced at scan time. For applications that don't require training, several general text reading engines can be used.

BSCAN's pattern recognition engine will automatically identify forms and determine their orientation and register them as well. In addition, it will recognize document separation marks on the first page of each document, thus eliminating the need for distinct separator pages.

For image cleanup and processing, Image Access has augmented existing technologies including a crop and deskew function that handles images produced from microfilm and from scanners that produce a black background around the document.

Production Customizations

While a completely pre-defined (i.e. non-customizable) software package is not likely to offer the highest performance operation for a typical capture application, building the ideal system from a toolkit can be a very difficult task. No matter how much time is spent on the drawing board, changes will be likely. If those changes can often be made while working with a scan or post-processing operator, the immediate feedback can ensure optimum performance. The result is dramatically reduced design time.

The first level of customization simply involves specifying screen layout (e.g. location and number of document view areas and location and content of control, status and data entry areas. This involves little more than specifying index field setup and which commands and functions are to be used.

The second level of customization typically includes some special scripting to improve the speed of indexing and/or the selection of high-speed image cleanup functions.

Beyond the second level, complex applications with many screens and modes of operation can be configured. The more sophisticated user can leverage his talents with the scripting facility and its rich set of index and image manipulation functions and a full compliment of conditional logic. There is no need to learn a new development environment or carry one from station to station when working on enhancements.

BSCAN scripts are built from a library of powerful functions that eliminate most of the discrete logic and algorithms that a C++ or VB application normally includes. For example, a single-line function is used to perform a correlation of an index field with a database and return numerous values. With the addition of a single parameter to a dialog box definition, a correlation is performed with every keystroke, offering immediate feedback to the data entry operator.

Batch Scanning with BSCAN

BSCAN is optimized for scanning and processing in batches, although it can be configured to release a document or even a page at a time. While live performance statistics can be monitored, BSCAN promotes monitoring and management by batches. The operator has the opportunity to take as long as necessary to check a batch before passing it to the next station or final release. Scanning in batches affords opportunities to minimize the load on a central control program such as the BSCAN Project Manager.

While BSCAN is managing the images, they are stored on standard operating system mass storage volumes, thus taking advantage of the low cost of high capacity, high performance hard disk drives for PCs. Raid storage is typically not necessary, since the most that can be lost is a single batch, and in the unlikely event of a system or disk crash, the batch can be rescanned and processed.

Large-Scale Image Capture Applications

Configurations requiring multiple BSCAN stations for scanning and processing can be setup to run independently, in a coordinated fashion or centrally controlled by the BSCAN Project Manager. Through the BSCAN Project Manager, the production stations can be configured to notify the Manager upon completion and release of a batch of scanned and/or processed images and then request the next work to do. While scanning a batch, the station can operate independently of servers or other stations and a virtually unlimited number of stations can be installed and managed from a single PC.

Versatile Workflow Options for Optimal Performance

Each BSCAN station is fully capable of performing any or all BSCAN functions, whether it is dedicated to scanning, indexing, QC or post-processing. Essentially, the only difference between the various scanning and non-scanning modules is price. Hence, if it makes sense to perform an indexing or image processing function at scan-time, it can easily be moved from a post-processing station to a scan-station.

Once a system has been in use for a few days or weeks, it may become apparent that one station is overloaded while another is running below capacity. It is a simple matter to reconfigure the stations so that they can swap or even share a process.

Production Management, Statistics and Reporting

All project, application, job, user, image and index data as well as performance statistics associated with any image capture process are available to BSCAN's scripting facility. Reporting can be performed periodically and/or triggered by an event such as starting and ending batches, documents, etc. Data can be graphically represented in real time using standard spreadsheet programs.

System administrators and technicians can retrieve hardware utilization statistics while production managers can receive operator performance information. If a performance problem is suspected and the standard reporting information is insufficient to pinpoint the problem, a temporary report can be created.

BSCAN Project Manager, a separate BSCAN module, provides a central point for setting up projects, assigning users and tasks and managing workloads.

 

Question or suggestion?
All Content © 2005 Image Access, Inc. All rights reserved.