Evaluation of model-based retrieval effectiveness with OCR text

Abstract We give a comprehensive report on our experiments with retrieval from OCR generated text using systems based on standard models of retrieval. More specifically, we show that for a set of queries, the average precision and recall is not affected by OCR errors across these systems. We also demonstrate that the ranking and feedback methods associated with these models are generally not robust enough to to deal with OCR errors. It is further shown that the OCR errors and garbage strings generated from the mistranslation of graphic objects increase the size of the index by a wide margin