Computers in Cardiology 1997
DOI: 10.1109/cic.1997.648140
|View full text |Cite
|
Sign up to set email alerts
|

A database for evaluation of algorithms for measurement of QT and other waveform intervals in the ECG

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
2
1

Citation Types

3
309
0
5

Publication Types

Select...
4
3
2

Relationship

0
9

Authors

Journals

citations
Cited by 420 publications
(331 citation statements)
references
References 3 publications
3
309
0
5
Order By: Relevance
“…Databases like the CCDB provide a varied source of raw data for those interested in developing new algorithms for reconstruction, segmentation, visualization, or image analysis of electron tomographic data. The utility of databases for spurring analysis tools has already been shown (Laguna et al, 1997;Shindyalov and Bourne, 2001).…”
Section: Resource For the Tomography Communitymentioning
confidence: 99%
“…Databases like the CCDB provide a varied source of raw data for those interested in developing new algorithms for reconstruction, segmentation, visualization, or image analysis of electron tomographic data. The utility of databases for spurring analysis tools has already been shown (Laguna et al, 1997;Shindyalov and Bourne, 2001).…”
Section: Resource For the Tomography Communitymentioning
confidence: 99%
“…Fig. 3, for example, shows the results for the sel103 recording of the MIT-BIH QT database [7]. The rMSDD and pNN50 have been calculated in a moving window of 100 beats width.…”
Section: Resultsmentioning
confidence: 99%
“…We evaluated the true positive rate and positive predictivity of our CNN model by comparing the resulting annotations with the reference annotations given for the QT database [4]. A strict tolerance level of 25 ms and 50 ms was set for the time difference between both annotations to count as successful (true positive).…”
Section: Resultsmentioning
confidence: 99%
“…were taken from the QT database [4]. We included realistic noisy segments, which were generated by applying the WFDB function nst [5] to clean records at different and very low signal-to-noise ratios (see [6]).…”
Section: Methodsmentioning
confidence: 99%