AI Article Synopsis
- This study assessed the accuracy of cloud-based speech recognition (SR) APIs for recognizing medical terms in Korean using real doctor-patient conversations.
- Naver Clova SR achieved the highest accuracy rate (75.1%), outperforming Google Speech-to-Text (50.9%) and Amazon Transcribe (57.9%) in recognizing medical terminology.
- While Naver Clova SR was most effective overall, the accuracy for longer words was similar across APIs, indicating room for improvement in medical terminology recognition technology.
Article Abstract
Background: There are limited data on the accuracy of cloud-based speech recognition (SR) open application programming interfaces (APIs) for medical terminology. This study aimed to evaluate the medical term recognition accuracy of current available cloud-based SR open APIs in Korean.
Methods: We analyzed the SR accuracy of currently available cloud-based SR open APIs using real doctor-patient conversation recordings collected from an outpatient clinic at a large tertiary medical center in Korea. For each original and SR transcription, we analyzed the accuracy rate of each cloud-based SR open API (i.e., the number of medical terms in the SR transcription per number of medical terms in the original transcription).
Results: A total of 112 doctor-patient conversation recordings were converted with three cloud-based SR open APIs (Naver Clova SR from Naver Corporation; Google Speech-to-Text from Alphabet Inc.; and Amazon Transcribe from Amazon), and each transcription was compared. Naver Clova SR (75.1%) showed the highest accuracy with the recognition of medical terms compared to the other open APIs (Google Speech-to-Text, 50.9%, < 0.001; Amazon Transcribe, 57.9%, < 0.001), and Amazon Transcribe demonstrated higher recognition accuracy compared to Google Speech-to-Text ( < 0.001). In the sub-analysis, Naver Clova SR showed the highest accuracy in all areas according to word classes, but the accuracy of words longer than five characters showed no statistical differences (Naver Clova SR, 52.6%; Google Speech-to-Text, 56.3%; Amazon Transcribe, 36.6%).
Conclusion: Among three current cloud-based SR open APIs, Naver Clova SR which manufactured by Korean company showed highest accuracy of medical terms in Korean, compared to Google Speech-to-Text and Amazon Transcribe. Although limitations are existing in the recognition of medical terminology, there is a lot of rooms for improvement of this promising technology by combining strengths of each SR engines.
Download full-text PDF |
Source |
|---|---|
| http://www.ncbi.nlm.nih.gov/pmc/articles/PMC9091429 | PMC |
| http://dx.doi.org/10.3346/jkms.2022.37.e144 | DOI Listing |
Publication Analysis
Top Keywords
Similar Publications
Characterization of pharmaceutical services in the prison system: cross-sectional study.
Cien Saude Colet
January 2025
Programa de Pós-Graduação em Assistência Farmacêutica, Universidade Federal do Rio Grande do Sul. R. São Luís 150, Santana. 90620-170 Porto Alegre RS Brasil.
Article Synopsis
- The study assessed the organization and infrastructure of pharmaceutical services in the prison system of Rio Grande do Sul, identifying major deficiencies in the primary health care framework.
- Most prison units lacked designated pharmacy areas and standardized procedures, with 78.4% reporting no operational guidelines for pharmacy activities.
- Only a small percentage had a pharmacist on site, and many did not adequately monitor medication storage conditions or have proper medication dispensing processes in place, highlighting significant weaknesses in the system.
Statistical analysis plan for hemodynamic phenotype-based, capillary refill time-targeted resuscitation in early septic shock: the ANDROMEDA-SHOCK-2 randomized clinical trial.
Crit Care Sci
January 2025
Departamento de Medicina Intensiva, Facultad de Medicina, Pontificia Universidad Católica de Chile - Santiago, Chile.
Background: ANDROMEDA-SHOCK 2 is an international, multicenter, randomized controlled trial comparing hemodynamic phenotype-based, capillary refill time-targeted resuscitation in early septic shock to standard care resuscitation to test the hypothesis that the former is associated with lower morbidity and mortality in terms of hierarchal analysis of outcomes.
Objective: To report the statistical plan for the ANDROMEDA--SHOCK 2 randomized clinical trial.
Methods: We briefly describe the trial design, patients, methods of randomization, interventions, outcomes, and sample size.
Assessment of the content and quality of YouTube videos on retinopathy of prematurity: a cross-sectional study.
Arq Bras Oftalmol
January 2025
Department of Ophthalmology, Tinaztepe University Hospital, Izmir, Turkey.
Purpose: This study aimed to evaluate the quality and reliability of YouTube videos as an educational resource about retinopathy of prematurity.
Methods: Videos were sourced from YouTube using the search terms "retinopathy of prematurity" and "premature retinopathy" with the default settings. Each video was assessed on the following metrics: views, likes, dislikes, comments, upload source, country of origin, view ratio, like ratio, and video power index.
Isolation of Human BAMBIhighMFGE8high Umbilical Cord-Derived Mesenchymal Stromal Cells.
J Vis Exp
January 2025
Department of Rheumatology and Immunology, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing University;
Umbilical cord-derived mesenchymal stromal/stem cells (UC-MSCs) present low immunogenicity and potent immunomodulatory effects for treating various diseases. Human UC-MSCs are a heterogeneous population consisting of three main subpopulations with different cell shapes, proliferation rates, differentiation abilities, and immune regulatory functions. Previously, BAMBIMFGE8 UC-MSCs, the first subgroup successfully isolated from UC-MSCs were found to fail to alleviate lupus nephritis.
View Article and Find Full Text PDFStair-descent phenotypes in community-dwelling older adults determined using high-level balance tasks.
Aging Clin Exp Res
January 2025
Department of Physical Medicine and Rehabilitation, Kansai Medical University, Osaka, Japan.
Background: Falls on stairs are a major cause of severe injuries among older adults, with stair descent posing significantly greater risks than ascent. Variations in stair descent phenotypes may reflect differences in physical function and biomechanical stability, and their identification may prevent falls.
Aims: This study aims to classify stair descent phenotypes in older adults and investigate the biomechanical and physical functional differences between these phenotypes using hierarchical cluster analysis.
Want AI Summaries of new PubMed Abstracts delivered to your In-box?
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!