Interactive archive card index conversion and verification

Museums and libraries have an urgent need to find ways of computerising typewritten archive card indexes to support computational analysis and automated database interrogation of their records. Existing commercial OCR software is however not capable of accurate archive conversion of this legacy typescript without significant tuning. Furthermore, external access patterns to future electronic archives are expected to be erratic and unpredictable, making full manual verification of electronic archives unjustifiably expensive. The paper therefore proposes a novel interactive verification approach for cost effective archive conversion. Efficient dictionary and string matching components are needed to enable interactive content indexing of archive card data; we describe possible implementations of these components which are currently under investigation. The proposed system will be deployed via the Internet to allow interactive access and content verification on-demand from anywhere in the world