MINING EVIDENCE AND SENSITIVE DATA USING COMMAND-LINE OCR
This talk will explore the use of open source command-line-based OCR tools to analyze images for sensitive/contextually-relevant information. This is especially relevant when faced with many images - too many to view manually - that might contain screenshots, faxes, scans, and other images-based files that may contain text important to an investigation or event. I will also discuss the shortcomings of the approach (as of today), areas for further research and improvements, and other applications of the methods. In addition, I will present a script written to automate much of the effort involved in using these methods.
Dennis Kuntz loves to break things to see what's inside and to figure out how to turn that into better ways to protect. He loves to challenge conventional wisdom and looks for ways to push the envelope and question "what is" in ways to help him and others grow.