When you index images into IDOL Server through CFS, KeyView can extract metadata from the images but cannot process their content. To enrich documents that represent images, you can send the images to an IDOL Image Server. The Image Server processes the images, and extracts useful information. CFS can run any analysis task that you have configured your Image Server to complete. For example, if the image is an image of text, you can run Optical Character Recognition. If the image contains faces, you can run face recognition. CFS takes the information extracted by Image Server and adds it to the document's metadata or content.
For example, you might want to tag all images that contain your company logo. After this information has been added to the document, it can be used by other IDOL operations.
Before you set up Image Analysis in CFS, complete the following steps.
After completing these steps, you can configure CFS. You can configure CFS to run analysis on all images, or you can configure CFS to process specific files.
To automatically run analysis on all images ingested by CFS, run the Lua script ImageAnalysis.lua
. This script is supplied with CFS, in the scripts
folder. This script runs analysis on all images that Image Server can process (including PDF files). By default, the script runs all the analysis tasks that have been configured in your Image Server configuration file.
In the [ImageServerSettings]
section of your CFS configuration file, you must specify the host name (or IP address) and ACI port of your Image Server.
The following example shows how to configure image analysis in the CFS configuration file:
[ImportTasks] Pre0=Lua:scripts/ImageAnalysis.lua [ImageServerSettings] ImageServerHost=server ImageServerPort=18000
In the [ImageServerSettings]
section, you can also specify the configuration parameters ImageAnalysisTaskSections
(to run only named analysis tasks) and ImageAnalysisTransform
(to specify an XSL transformation to use to transform the metadata returned from Image Server, before it is added to the document). For more information about these parameters, refer to the CFS documentation.
To run analysis on specific images, you must use the ImageServerAnalysis
import task, and write your own Lua script to identify the documents that you want to process. The ImageServerAnalysis
task processes only documents that have the field AUTN_NEEDS_IMAGE_SERVER_ANALYSIS
, so your script must add this field to any document that you want to process.
The following example shows how to configure the ImageServerAnalysis
task in the CFS configuration file:
[ImportTasks] Post0=Lua:EnableImageAnalysis.lua Post1=ImageServerAnalysis:ImageServerSettings [ImageServerSettings] ImageServerHost=server ImageServerPort=18000 ImageAnalysisTaskSections=dococrtask,facerectask,objecttask ImageAnalysisTransform=xslt/My_Transform.xsl
The Post0
import task runs your Lua script, which must determine whether a file is suitable for image analysis, and add the field AUTN_NEEDS_IMAGE_SERVER_ANALYSIS
to any document that you want to process.
The Post1
import task is the ImageServerAnalysis
task. This parameter specifies the name of a section in the CFS configuration file that contains the settings for image analysis.
In the named section, specify the host name (or IP address), and ACI port of your Image Server. You can also specify the configuration parameters ImageAnalysisTaskSections
(to run only named analysis tasks) and ImageAnalysisTransform
(to specify an XSL transformation to use to transform the metadata returned from Image Server, before it is added to the document). For more information about these parameters, refer to the CFS documentation.
You can include image analysis as part of a custom Lua script by using the Lua functions provided by CFS. CFS provides the functions analyze_image_in_document
and analyze_image_in_file
. For more information about these functions, refer to the Connector Framework Server Administration Guide.
|