A Non-Visual Photo Collection Browser based on Automatically Generated Text Descriptions

Abstract

This study presents a textual photo collection browser that automatically and quickly analyses large personal photo collections and produces textual reports that can be accessed by blind users using either text-to-speech or Braille output devices. The textual photo browser exploits recent advances in image collection analysis and the strategy does not rely on manual image tagging. The reports produced by the textual image browser gives the user a gist about where, when and what the photographer was doing in the form of a story. Although yet crude, the strategy can give blind users a valuable overview about the contents of large image collections and individual images which otherwise are totally inaccessible without vision.