CapnoBase is a collaborative research project that provides an online database of respiratory signals and labels obtained from capnography, spirometry and pulse oximetry.

The CapnoBase Dataverse ( currently contains 6 datasets with annotated respiratory signals such as inhaled and exhaled carbon-dioxide (CO2) also known as capnogram, respiratory flow, pressure and photoplethysmogram (PPG) obtained from anesthesia monitors during elective surgery.
The database includes two benchmark datasets: One for detection of respiratory events (CapnoBase Respiratory Event Benchmark) and another for extraction of respiratory rate from PPG (CapnoBase IEEE TBME Respiratory Rate Benchmark). These benchmarks are widely used to validate and compare algorithms in scientific literature.
The "CSL Pulse Oximetry Artifact Labels" contains the artifact labels created for the external Complex System Laboratory (Portland State University) Pulse Oximetry (infrared) Beat Detection Benchmark available here.
Before the creation of CapnoBase in 2010, there was no benchmark dataset publicly available for respiratory signal analysis. A benchmark dataset is required to objectively assess and compare algorithm performance and share the results without the need to disclose in detail the source code of the algorithms (although this would be desired!). Since January 2021, CapnoBase is hosted on

File Formats

The file formats used by CapnoBase are comma separated values (.csv / .tab) and proprietary Matlab (.mat) files.
Data types, such as signals, metadata, labels (annotations), comments and parameters, are stored in separate files, organized by recording.
The coding of filenames is as follows:
e.g. A66_meta.csv or 100B_signal.mat
The CapnoBase Dataverse provides files to be downloaded in a compressed file format (.zip) or as individual files.

Folder and file structures

Each dataset in the dataverse is organized into subfolders. "preview/" contains figures of the waveforms (capnogram and/ or PPG) that can be found in the dataset. "data/" contains the all the datafiles. This folder is subdivided into a "csv/" and "mat/" folder containing the respective data formats.
The mat files store the data in a structured format.
In csv files, the structure fields are replaced by unique row headers.
E.g. If the matlab stucture is meta.sensor.co2.manufacturer, the corresponding csv row header would be meta_sensor_co2_manufacturer.

Requesting access

For certain data you will need to request access. For this you have to create an account and agree to general terms and conditions.
After being logged in, 1) go to the dataset of choice, 2) select all files of interest (use the selection box in the header to select all in this dataset), 3) click "Request Access" button in header. You will now gain access within 1-2 working days, often faster.
Please do not send emails or other contact requests to get access to files. The above procedure is the only way to gain access.


If you are using CapnoBase material (data, information and software from this website) for your work, please cite our original abstract that was presented at the 2010 Society for Technology in Anesthesia Annual Meeting. If you are using the PPG signals (respiratory rate benchmark), the IEEE TBME 2013 paper can be cited.
If you are preparing a manuscript for publication, please add the following citations to your bibliography. If you contact us after publication, we can update the citation list.
  • Karlen, W. and Raman, S. and Ansermino, J. M. and Dumont, G. A. (2013). Multiparameter respiratory rate estimation from the photoplethysmogram. IEEE transactions on bio-medical engineering. 60 (7), 1946-53. 10.1109/TBME.2013.2246160
  • Karlen, W. and Turner, M. and Cooke, E. and Dumont, G. and Ansermino, J. M. (2010). CapnoBase: Signal database and tools to collect, share and annotate respiratory signals. 20.500.11850/87887


The CapnoBase was developed on the initiative of Dr. Walter Karlen and Dr. Mark Ansermino at the Electrical and Computer Engineering in Medicine research group (ECEM) at the University of British Columbia, Vancouver, Canada between 2009 and 2010.
We are grateful to all contributors, in particular:
  • Data preparation and analysis have been performed by the Pediatric Anesthesia Research Team at the BC Children's Hospital Research
  • Data recording was performed at the St. Pauls Hospital, Vancouver and the British Columbia Children's Hospital, Vancouver

General Terms of Use

Please check individual dataset Terms of Use for additional details
  • The databases and the software on are available free of charge for students and researchers.
  • All published work based on any information, data, software, or other materials from should include a citation of the CapnoBase abstract presented at the 2010 Society for Technology in Anesthesia Annual Meeting and the IEEE TMBE 2013 paper (see "Literature" above and the literature references in the individual datasets).
  • It is not permitted to copy or redistribute any data and/or software from / dataverse without citing as the source and/or including all license agreement/terms of use.
  • All data in the CapnoBase Dataverse has been anonymized and meets ethical guidelines for data privacy protection. Any type of work on re-identification of subjects is not allowed.
  • Local institutional Research Ethics Board approval has been obtained for recording the CapnoBase Dataverse data. When using CapnoBase Dataverse data, it is recommended that researchers apply for Research Ethics Board approval at their own institution for reuse of the data, by citing CapnoBase.
  • By using the CapnoBase Dataverse or any associated applications, you are agreeing to do so at your own risk, and, the Electrical & Computer Engineering in Medicine (ECEM) group or any other person involved with this project are not liable for any loss or corruption of data, or any other problems resulting from the use of CapnoBase software or data.
  • To access all data in the password protected part of the CapnoBase Dataverse, users have to sign in and agree with the end-user license agreement. To request a whole data set and not a single file, please then select all documents and request access with the matching button.
Featured Dataverses

In order to use this feature you must have at least one published dataverse.

Publish Dataverse

Are you sure you want to publish your dataverse? Once you do so it must remain published.

Publish Dataverse

This dataverse cannot be published because the dataverse it is in has not been published.

Delete Dataverse

Are you sure you want to delete your dataverse? You cannot undelete this dataverse.

Advanced Search

1 to 6 of 6 Results
Jan 3, 2021
Karlen, Walter, 2021, "CapnoBase InVivo Dataset",, Borealis, V1, UNF:6:IoEbUHgJExSfNXTnhZ64bw== [fileUNF]
The in-vivo dataset contains capnography signals that were recorded during real clinical cases. Information that could possibly identify the source of the data has been removed.
Jan 3, 2021
Karlen, Walter, 2021, "CapnoBase Respiratory Event Benchmark",, Borealis, V1, UNF:6:qUQWh7sXzT6UWsNkLfE50g== [fileUNF]
The CapnoBase benchmark dataset contains 44 scenarios that were recorded in the same manner as the in-vivo dataset. The scenarios contain very typical capnography and spirometry signals or patient conditions that may arise during anesthesia. The benchmark dataset is used by resea...
Jan 3, 2021
Karlen, Walter, 2021, "CapnoBase Simulation Dataset",, Borealis, V1, UNF:6:dBeFZTVT2s+UCDhOQpCpSw== [fileUNF]
The simulation dataset was produced with a computer model that simulates the behavior of the human cardio-respiratory system. Producing artificial capnography signals is useful when signals that are otherwise hard to obtain are required, or when the condition or change in state o...
Jan 3, 2021
Karlen, Walter, 2021, "CapnoBase 8-minute (long) Dataset",, Borealis, V1, UNF:6:jFmr4FBCVm5bFFhw6FlApA== [fileUNF]
This data set contains currently nine longer recordings of 8 minutes. All are pediatric cases are spontaneously breathing. The cases have not been rated for respiratory events, but they contain additionally ECG and photoplethysmogram recordings.
Jan 3, 2021
Karlen, Walter, 2021, "CapnoBase IEEE TBME Respiratory Rate Benchmark",, Borealis, V1, UNF:6:V0wQVsilTA0RcI09PzLJdw== [fileUNF]
The CapnoBase TBME RR benchmark dataset contains 42 cases of 8-min recordings. In addition to the CO2 waveforms (capnograms), these cases have also the Photoplethysmogram (PPG) from pulse oximetry available. Labels from an expert are available for pulse peaks from PPG and breaths...
Jan 3, 2021
Karlen, Walter, 2021, "CSL Pulse Oximetry Artifact Labels",, Borealis, V1
This dataset contains the beat and artifact labels created for the comparison of beat detection algorithms for photoplethysmogram analysis as published in Karlen W, Ansermino JM, Dumont GA. "Adaptive Pulse Segmentation and Artifact Detection in Photoplethysmography for Mobile App...
Add Data

Sign up or log in to create a dataverse or add a dataset.

Share Dataverse

Share this dataverse on your favorite social media networks.

Link Dataverse
Reset Modifications

Are you sure you want to reset the selected metadata fields? If you do this, any customizations (hidden, required, optional) you have done will no longer appear.