I did some experimentation wrapping the Perl Image::ExifTool module (along with Image::OCR::Tesseract) in some code that exposed it as a SOAP service for use in a Fedora Commons ingest service. It seemed to work well enough for bulk file processing in testing, though the approach of a custom ingest system, in general, was eventually abandoned when consultants were brought in.
Were I to do it again I'd probably also add a REST interface to the generic service wrapper.
From: Code for Libraries [mailto:[log in to unmask]] On Behalf Of Edward Summers
Sent: Tuesday, December 17, 2013 4:54 PM
To: [log in to unmask]
Subject: Re: [CODE4LIB] Automated Embedded Metadata Extraction in Photographs: Possible or Pipedream?
I remember hearing somewhere that ExifTool is pretty good for extracting image metadata.