
Figure 2-5 OCR engine
The figure above shows the process flow for OCR processing in DSS. When DSS receives a job that
requires OCR processing, it invokes the I.R.I.S. OCR engine using COM (Component Object Model).
The image data/document is transferred together with control parameters, such as the required
output file type. Once OCR processing is completed, the searchable document is passed back to
DSS which delivers the document to the destination.
DSS is a multi-threaded application and will launch multiple instances of the OCR engine when there
are multiple jobs in the queue that require OCR processing. We refer to this as ‘parallel processing of
OCR jobs’. This makes the OCR feature scalable, which means that average job processing times
will be improved if the server's resources are improved. For instance, adding additional CPUs and
more memory to the server will improve the average processing time of each OCR job when the
server is processing multiple jobs simultaneously. This is a significant improvement over previous
versions of DSS, where OCR processing was serial.
Database
DSS uses Microsoft SQL Server 2008 SP3 Express Edition to host the DSS database. The database
is used to hold job logs, address books, event logs, and some configuration data.
Microsoft SQL Server 2008 is a database management system (DBMS). Within the DBMS, DSS
creates two databases for specific use by DSS, named as follows:
●
DSS_Customer
●
DSS_Machine
The SQL Server 2008 database instance name is “HPDSS2008.”
It is possible to configure DSS to use a DBMS other than Microsoft SQL Server 2008 SP3 Express
Edition. If a different DBMS is specified during installation, DSS will not install SQL server on the local
server. Even if the local database that DSS installs is used at first, the system can be configured later
to use a different DBMS, but some data will be lost during the switch.
ENWW Components 15
Commentaires sur ces manuels