Speech-controlled generation of radiology reports.

摘要: Voice entry has been successfully employed to generate radiology reports with a word recognizer 1,000-word lexicon capacity. About 50% of were able be dictated single 900-word lexicon. This was split into five sections by anatomic or subspecialty application. Each augmented 900 words. By switching from one another, it possible dictate more than 70% reports. With exclusive use three lexicons in areas (gastrointestinal radiology, neuroradiology, and mammography), further modification the respective vocabulary, employ system 88% time. Twelve percent cases included wording that beyond scope Computer subsets allow different translations some words when used contexts have used. Some these are as triggers will print whole lines, sentences, even complete Dictation times voice take about 20% longer. Recognition reliability greater 95%.