Figure 4
From: An intelligent listening framework for capturing encounter notes from a doctor-patient dialog
Proof-of-concept output. A sample run of our proof-of-concept system, showing the marked-up text (right) and the key for the mark-up (left). Notice that despite some false positives, many words and phrases are detected properly. The F-measure of detection accuracy is 73.6%. Additionally, 85.7% of the correctly identified items are also categorized correctly. Note that this is not meant to be an example of actual physician speech, but a combination of technical and non-technical medical terms in casual speech.