How to use multiple languages and OCR engines in the same ZySCAN job

SYMPTOMS

This article describes the options available to use multiple languages and OCR engines in the same ZySCAN job.

CAUSE

Not applicable.

RESOLUTION

To detect multiple languages and use different OCR engines in a ZySCAN job, the Use multiple languages/engines option can be selected in the ZyOCR configuration stage of the ZySCAN template configuration wizard.

• Launch ZySCAN.
• From the Template menu, select New Template.
• Define new job template or modify an existing one.
• Proceed with the Template configuration wizard until the ZyOCR configuration stage.
• From the Languages tab, enable the Use multiple languages/engines option.
• When Store all output option is selected, OCR for all selected languages will be stored as metadata of the document.
• When Select best result option is used instead, only the language that most resembles the text will be stored. In this case, when more than one language has been added to the list, the Mode option becomes available. Find more details about this option below.
• Two extra index fields will be created: Language_Code and Language_Name.
• Proceed with the other Template configuration wizard steps as appropriate till completion.

Mode: Per job
When all of the documents in the same job are expected to have been written in the same language, Mode can be set to Per job. The language detected in the first document will be used for all documents.

Mode: Per document and Per Page
When the language is expected to be different for each document or each page within documents, Mode can be set to Per document or Per page. Processing time will increase for this level of detection, but the language that most resembles the text in each document or each page will be stored.

NOTE: Please note that this configuration will only take effect if enabled from the Template configuration wizard. Changing the same ZyOCR Stage settings while a job is being processed will not create the extra index fields, and will not detect the languages used in the documents.

APPLIES TO

3.6