The work describes methods used in a laboratory order form recognition module of a hospital information system. Three-level form analysis architecture is proposed. The lower alphabetical level is responsible for separate character recognition. On the intermediate level, recognised strings are verified against the lexicons of items specific for a particular form field. Probabilistic model is used to select the set of most probable items. On the upper level, the dependencies between the form data items are taken into account to further improve the recognition performance. The presented approach was implemented in the medical information system supporting clinic laboratory operation. The laboratory test orders prepared manually by the physician in the paper form, in the net of distributed outpatient clinics are processed in the central hospital laboratory. In the central laboratory the paper forms are scanned, recognised and entered into the information system. The performance tests results are discussed and some further improvements of the applied recognition method are also suggested in the paper.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.