Analyze Images

CFS extracts metadata from image files but cannot process their content, so by default documents that represent images are indexed without any content.

To enrich documents that represent image files, you can send the images to an IDOL Image Server for analysis. Image Server can return information about the content of an image, for example:

You can run any analysis task that you have configured in your Image Server configuration file. For more information about the analysis tasks that you can run, refer to the Image Server Administration Guide. Some of the analysis tasks require you to train Image Server before processing images.

To send images to an Image Server use the ImageServerAnalysis import task. This task uses the analyze action provided by Image Server.

Note: The ImageServerAnalysis task requires Image Server 10.3 or later.

CFS adds the results of image analysis to the document. By default, the information is added in the same structure that is returned from Image Server but you can use an XSL template to transform the information before CFS adds it to the document. CFS includes a sample XSL transformation, ImageServerAnalysisTransform.xsl, that produces output similar to the following example:

<DOCUMENT>
   <DREREFERENCE>http://www.example.com/</DREREFERENCE>
   <AUTN_NEEDS_IMAGE_SERVER_ANALYSIS/>
   <FOUND_LOGO>Detected object</FOUND_LOGO>
   <RECOGNIZED_FACE>Detected face</RECOGNIZED_FACE>
   <DRECONTENT>Text extracted from the image using OCR</DRECONTENT>
</DOCUMENT>

_HP_HTML5_bannerTitle.htm