Document Detective - 3.0

Document Detective is an electronic document review and sanitization tool that meets the intent of the Government's "100% reliable human review" criteria for transferring Microsoft Office files, Adobe PDF files, and text files across security boundaries in their native format. Document Detective runs as a desktop application on an individual workstation, and has one interface for rigorous document reviews and another for easy sanitization. Document Detective has been approved and is being used at several Government and contractor facilities for Trusted Downloading.

The review interface organizes the document contents into a hierarchical browser view similar to the Microsoft File Explorer. Objects are organized into folders in the left window, and the contents of an object are shown on the right. Document Detective analyzes the document for indicators that information might be obscured from the user's view, and it marks those items with a red warning dot. Document Detective looks for cropping, resizing and other geometry indicators that could indicate hidden data. All text is checked with a regular expression keyword scanner for words that could indicate the presence of sensitive or classified information.

The Flatten Document interface can remove much of the document's hidden data with just a few mouse clicks. Depending on the security level selected, the Flatten Document function automatically converts dangerous OLE objects into images, and reprocesses images such that the resolution is 200 dpi, as displayed with any cropped areas permanently removed from the image. The Flatten Document interface is very aggressive, and may alter the document's appearance. Document Detective provides a side-by-side review capability that lets the user compare the 'flattened' document's view with the original document. Problems introduced by the flattening process can be quickly identified and remedied in this view. A special toolbar lets the user scroll through both documents in tandem, and allows the user to accept or reject the results with the click of a button.