Improving language models for radiology speech recognition
| Autor: | Curtis P. Langlotz, John M. Paulett |
|---|---|
| Rok vydání: | 2009 |
| Předmět: |
Diagnostic Imaging
medicine.medical_specialty Medical Records Systems Computerized Computer science Speech recognition Health Informatics computer.software_genre Speech Recognition Software Artificial Intelligence Medical imaging medicine Humans Analysis of Variance Modality (human–computer interaction) business.industry Computer Science Applications Word lists by frequency Radiology Information Systems n-gram Language model Radiology Affect (linguistics) Artificial intelligence business Productivity (linguistics) computer Natural language processing |
| Zdroj: | Journal of Biomedical Informatics. 42:53-58 |
| ISSN: | 1532-0464 |
| DOI: | 10.1016/j.jbi.2008.08.001 |
| Popis: | Speech recognition systems have become increasingly popular as a means to produce radiology reports, for reasons both of efficiency and of cost. However, the suboptimal recognition accuracy of these systems can affect the productivity of the radiologists creating the text reports. We analyzed a database of over two million de-identified radiology reports to determine the strongest determinants of word frequency. Our results showed that body site and imaging modality had a similar influence on the frequency of words and of three-word phrases as did the identity of the speaker. These findings suggest that the accuracy of speech recognition systems could be significantly enhanced by further tailoring their language models to body site and imaging modality, which are readily available at the time of report creation. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect