The graphical interface BioSigBrowser was born in 2007 and it is continuously updated by the BioSignal Interpretation and Computational Simulation Group (BSICoS group) located in Zaragoza (Spain).   The main task of this interface is to provide researchers, working in the cardiovascular field, advanced tools for signal processing developed by BSICoS group.   BioSigBrowser is developed under Matlab® language.  BioSigBrowser is structured allowing the integration of new modules easily.  Furthermore, methods included in the graphical interface are available for parallel computing.   Although, most of the methods were developed for assessing the ECG characteristics, advance processing tools may be suitable for other signals.

BioSigBrowser graphical interface allows visualizing your signals friendly.  But not only signals, series extracted from annotations made during the recordings or extracted from signal characteristics.

Signal Window of BioSigBrowser



How to start?

First of all, we should change the current directory of Matlab® to this one:


Once, we must not change the current directory in order to use BioSigBrowser properly.

Type “biosigbrowser” in the command window and all functionalities, that the interface is going to use, will be available because the correspondent paths will be added.  (The user should not modify paths manually after a BioSigBrowser session since when a session is finished all paths included are removed.)


Formats supported

BioSigBrowser supports:  MIT (WFDB format [physionet]), Mortara Bin Format, Matlab® files, Biopac MP System format and GDF (General Data Format).  Moreover, other formats (not checked) are readable by some routines provided by BioSig project (   Usually recordings are acquired with a fixed sample rate but the possibility of combining multiple sample rates depending on the necessity of each particular signal in the same file exists and this is also readable by BioSigBrowser.

Pre-processed signals by BioSigBrowser interface can be stored for further analysis in two formats nowadays: Matlab® files and MIT format.


Loading a recording

In the menu, “Actions” -> “Open recording”


Selecting recordings to open in BioSigBrowser


In this case, we chose WFDB file provided by physionet.  Although WFDB format consists in at least two files (*.dat & *.hea), the file is identified by the *.hea file.  For further details about WFDB format visit

BioSigBrowser informs you about the path where the results obtained by methods using the platform will be stored.  By default, this results directory will be created in the same folder as the recording with the name of “iBdir”.  This folder is not a temporary one.  Once the BioSigBrowser session is finished, the results will be still there.   This results folder can be changed by other one in “Actions” à “Change Temporary Directory”.

Afterwards, another window will appear whose name is “Select the nature of the leads”.  As mentioned in the introduction, most of the algorithms included in this interface are related with electrocardiogram signal processing.  Therefore, if our current recording consisted in electroencephalograms (EEG) or PPG or other sort of biomedical signal, it makes no sense that all ECG algorithms would be available, only will be those algorithms which are related with.

Selection of the nature of each recording lead



Saving a recording

BioSigBrowser contains several algorithms which are focused on preprocessing. Preprocessing is a natural stage to adequate raw signals to advanced signal processing algorithms. Nevertheless, not all advanced signal modules included in BioSigBrowser platform were built to be fed by this preprocessing signal. Therefore, preprocessing signal needs to be saved as a new recording to solve this problematic.


What does BioSigBrowser display? (Signal Window)

Signal Window for BioSigBrowser interface


  1. Menu Bar: It is divided into:
    1. Actions
    2. Edit
    3. View
    4. Pre-processing
    5. Analysis
    6. Post-processing
    7. Show
    8. Signal Info
    9. Batch Mode Assistant
    10. Help
  2. Shortcuts: trying to save your time with the most used steps.
  3. Information related with the current BioSigBrowser session is shown here.
  4. In this part the label of each signal is shown as well as a 1 mV scale as reference.
  5. The time onset of the row is shown. In case that more than one row was displayed, then each row would have its own time onset.
  6. This blue line displays the time scale in seconds. By default 10 seconds for each row is shown.
  7. User can use these edit boxes to browse at any time or sample on the recording.
  8. A slide bar is available to browse on the recording quickly.




By clicking with the right button of the mouse in the Signal Window a popup menu is shown. Among the possibilities offered there, user can change the row duration, the amplitude of the signals and the leads to be displayed. VCG loop representation is also included in popup menu.


Loading annotation file/s

Annotations are time marks that usually symbolize certain characteristics in the original signal. For instance, ECG main beat waves and their boundaries are considered as annotations which allow further advanced signal processing analysis. These annotations can be provided either by some BioSigBrowser algorithm or by external software or even by clinicians in certain databases.

Annotations are stored in files with a particular format. Formats supported are: MIT/WFDB format, and Matlab® format (designed for BioSigBrowser, for more information, see Annotations Format section).

To display annotations superimposed with the signal, follow “Actions” -> “Display annotation file/s”. By default, annotation files are searched in the results folder generated by the interface at the beginning of the session (remember that its name is “iBdir”).


Supposing we loaded an ECG recording and we used the ECG detector/delineator to extract beat waves and boundaries of these waves, then the annotation files generated would be in the “iBdir” folder or in the results folder if we had changed it.


Annotation file/s to be displayed superimposed in the signal.

More than one annotation file can be selected to be displayed.  Usually, each annotation file corresponds to a specific lead (Thinking of single lead detection).

Relationship between leads and annotation files to be displayed. The content of the annotation file loaded previously will be displayed in the lead selected by user or by auto-detection.

Selecting one of the annotation files, user can chose on which lead want to superimpose the annotations.  Auto-detection is available only if the annotations file name ends as the label of the lead which corresponds.

Annotations to be shown in the Signal Window.


Then, user selects the annotations to display.


Signal Window: annotations displayed on V1 lead.


Now the information related with the current BioSigBrowser session is updated and the loaded annotation file is shown in the top of the panel in the Signal Window.


Annotation/Events Management

Annotations and events are time marks or time intervals that can be produced by algorithms. In other sections, it was shown how to load files which contain annotations or events and how to display them superimposed in the signal.

However, annotations and events can be modified by user manually. This section deals with how to manage them, modifying, generating new ones or removing them.
The way to access the menu is by right clicking on the Signal Window:

First of all, it is worthy to note that annotations or events derived from a specific algorithm or given by an external clinician wont be modified. Then, new file will be generated with the new marks or modifications.


New annotations are considered as provisional marks. When one provisional mark is set, user will be asked to introduce a description (not be confused by the label of the annotation since all of them are labeled as provisional). Then, the mark appears in the Window Signal as a blue vertical line and at the top the description in yellow. To change the position of this mark just right clicking on the line or on the yellow description (it turns blue) and another menu is shown allowing: to remove and to change position and description.

Provisional marks can be stored liking them to an existing annotation file. If no annotation file is loaded, see “Loading annotation file/s”. Then user will be asked to link them to a type of mark.

In the case of modifying annotations loaded, the procedure is quite similar. Just right clicking on the bullets it will appear the menu. A new annotation file (named after the same as the original but ended with “aux”) is generated to protect original annotations.


As it happens with annotations, new events will be considered as provisional until they will be stored. The way to introduce an event is quite similar as annotations but with the difference of an event consists in two time locations. So, the first event will be the initial time and the second will be the ending (depending whether is located after the first one). Once both time locations are selected, an event was generated. As it is provisional, the vertical lines are painted in blue. Event time marks works as annotations, thus, they can be changed just right clicking on the line or in the description and the menu to manage them is displayed. It is worth to note that events are also drawn in Trend Window in which series are shown.

If no annotation file is loaded and user will store the events, then, a new file will be generated. Events are not loaded. To load events, go to “Actions” -> “Set time analysis” and select load events. Once this is done, then BioSigBrowser will ask user to show events. If showing events is in mode On, then events will be displayed in green, differentiating them from the provisional ones that remember they are drawn in blue.


Pre-processing methods


Pre-processing consists in a wide variety of algorithms. Basically, the main idea of pre-processing has to be shown as a previous step for the further analysis. These previous tasks can be: clean a certain contamination in the signal such as muscular noise or power line noise, detection of beats (in the case of ECG recordings), and so on. In the current version BioSigBrowser has available the following pre-processing modules:

  • Filters
  • ECG detector/delineator
  • Heart beat classifier
  • Other ECG leads (by linear combination of the recorded ones)
  • PPG Artefact detector
  • PPG Pulse detector
  • Principal Component Analysis (PCA)
  • Independent Component Analysis (ICA)
  • Periodic Component Analysis (pCA)
  • Segment Karhunen Loewe Transform
  • Ectopic beat detection
  • Series generation
  • Events generation by segmentation



A filter designer is provided by BioSigBrowser.   Three different types of filters are available: IIR, FIR and Adaptive Notch Filter.  Parameters of each filter type became enabling when filter is selected.  Furthermore, filter response is also provided in order to choose the better parameters for the filter designed by user.



Filter design interface in BioSigBrowser.

Besides, baseline wander removal can be done applying a technique based on using cubic splines.  It is based on ECG characteristics.  Segment PQ is isoelectric meaning it should be at 0 mV.  This characteristic can be profited in order to force the ECG to return in these particular segments at zero voltage.  The strategy proposed uses the time location of each fiducial beats in time, then, 55 ms are left-shifted to place the PQ point.  Afterwards, a reconstruction of a baseline is generated by using cubic splines and it is removed from the original signal achieving a nice baseline wander removal.


Baseline wander removal by using cubic splines on PQ segment of ECG signals.

ECG detector/delineator

Morphological characteristics of ECG (electrocardiographic) signal contain relevant information about electrical changes occurred in the heart.  Some of them are generated in the atria and others in the ventricles.  Looking at one single ECG beat it is easy to see at least three parts clearly distinguishable.


Principal segments and waves in a single heart beat electrical activity.

The first wave is named P wave which reflects the atrial depolarization.  Its presence in the beat is a symptom of the beat was generated in the sinus-atria node.  The second part is the qrs-complex.  As it is own name indicates it is composed of three waves, Q, R and S.  The underlying physiology of this complex is due to the ending of the atrial repolarization as well as the ventricles depolarization.  Moreover, the R wave peak is usually considered as the reference location of a heart beat usually named as fiducial point.  The final part of the heart beat consists of the repolarization of the ventricles and its representation on the ECG is registered at the T wave.

Each particular wave determined by their peaks and boundaries is crucial for advanced signal processing.   The better resolution either in detection (in case of peaks) or in the delineation (in the case of boundaries), the greater accuracy in the characterization of each physiological part in the ECG.  And it is suggested to check those marks generated once the automatic ECG detector/delineator was applied.

ECG Detector/Delineator included in BioSigBrowser are based on the wavelet transform.  Basically, it exploits the frequency content of each main wave decomposing them in different time-scale projections. Further details of accuracy and thresholds used by this algorithm can be found elsewhere {JP2004}.

As it was mentioned before, detection referrers to detect wave peaks whereas delineation means to find the boundaries of the waves.  And BioSigBrowser allows doing this by three ways.

  1. Single lead: The detection and delineation is done in a single lead using wavelet transform and using several sort of thresholds and time windows.
  2. Multi-lead: It is based on vectorcardiogram which is computed by three orthogonal leads. The loop, which is the graphical view of each orthogonal leads on the three axes, allows finding the maximum direction which is related with the projection of the electrical axes of the heart.  Then, a pseudo-lead can be generated as a projection of theses leads over this maximum direction and single lead approach can be applied to achieve one representative set of marks.  The marks obtained through this algorithm reflect the global electrical phenomena.  One of the advantages of this algorithm yields on the vectorcardiogram which is less affected by respiratory movements than single lead approach.  However, on the other hand, this ECG detection option is much time-consuming than single lead approach.

Single lead + rules: This third option combines the single lead approach, which is less time-consuming, and a set of post-processing rules.  Single lead approach is done over all desired leads and afterwards the post-processing rules are applied.  These rules are based on the consistency of marks meaning that they were grouped and as a result outliers are not considered.  Therefore, a unique set of marks looking for representing the whole electrical phenomena from the early mark from each single wave onset to the latest mark offset is obtained.

ECG Detector/Delineator setup interface.

External marks can be used for limiting the searching step of delineator.  These external marks, usually fiducial points, can be provided by some expert (clinician) or by other kind of ECG detector.

Heart beat classifier

Heart beat classifier included in BioSigBrowser uses feature selection driven by database generalization criteria.  It presents three operational modes: fully automatic, slightly assisted and fully assisted.  The number of clusters related with the different classes can be set by user as well.  Once all setting parameters are selected, a friendly interface guides user when operational mode is slightly or fully assisted.

Heart beat classifier setup interface.

Other ECG leads

ECG is registered by electrodes located in a certain body positions which they are well established and standardizes by electrophysiology.  Their positions in the body surface give us information about the projection of the electrical activity that is happening in the heart. Each electrode registers its own electrical activity named leads.  In this section, it is presented three ways of generate extra leads when ECG recording allows that.

ECG lead locations and their relations.

Dower transformation allows computing X, Y and Z leads (named Frank leads) as a linear combination of the standard leads (V1, V2, V3, V4, V5, V6, I, II and III).  These Frank leads are orthogonal synthetizing the electrical information coming out from the heart.

As well as Dower is one of the transformations, other methods or linear combinations have been proposed whose aims are achieving these Frank leads such as Levkov’s transformation or Kors regression.  All of them are available in the BioSigBrowser software.

In addition, if the leads recorded were X, Y and Z coming from, for instance EASI leads, the step back can be done by applying the inverse Dower transform.

Furthermore, extremity/limb leads can also be computed from standard leads.


PPG artefact detector (DAP detector)

Photoplestimografic (PPG) artefact detector was developed based on Hjorth parameters.  It exploits the fact of when a signal differs largely from an oscillatory signal then it is supposed to be an artefact.  More details of this algorithm are described elsewhere {EDU2008}.

PPG artefact detector setup interface.

PPG Pulse Detector

The pulse detector included in BioSigBrowser consists of two phases: firstly a linear filtering transformation and secondly an adaptive threshold operation.  The filtering transformation accentuates the abrupt upward slopes of the PPG pulses avoiding the false detection of irregular pulses.  For the second part, a time varying threshold gradually decreasing between detections is used.  Further details of the algorithm can be found elsewhere {Lazaro}.

Extracted from Lazaro et al (2013)

PPG pulse detector setup interface.

Principal Component Analysis (PCA)

Principal component analysis (PCA) is a transformation which allows converting a set of observations possibly correlated into a set of values linearly uncorrelated called principal components.  The transformation is defined in such a way that the new orthogonal space exploits the variance of the data.  First principal component will have the largest possible variance; the second will have the highest variance under the constraint that it is orthogonal to the first component and so on.

Setup to compute principal component analysis (PCA).Setup to compute principal component analysis (PCA).



The new basis can be generated based on: Time interval; Events; and percentage of variance.

  • Time interval: a piece of the signal will be used to generate the new eigenvectors and eigenvalues to project the original data.
  • Events: instead of a piece of signal, several time intervals corresponding with a certain characteristic (for instance, T wave interval) can be used to generate a transformation which exploits the similarity with the selected basis.
  • Variance: A percentage of variance is selected and then the algorithm generates the maximum number of new principal components that fits that variance.

Independent Component Analysis (ICA)

                This module is under construction.

Periodic Component Analysis (pCA)

                This module is under construction.

Segment Karhunen Loewe Transform

                The KLT is a signal-dependent linear transform that is optimal for any given number of transform parameters.  It concentrates the maximum signal information in the minimum number of parameters, and it defines the domain where the signal and noise are most separated.  A significant limitation of the KLT is that it is necessary to collect a representative “training” set of the signals to be analysed, in order to derive the KLT basis functions (eigenfunctions).  Therefore, the performance of the KLT depends on how well the training set has been constructed.

                The module included in the BioSigBrowser was used to determine the basis for ST-T segments specifically to be used in ischemia detection.

Ectopic beat detection

One ectopic beat is defined as the beat that occurs out of the sinus node.  One characteristic of these ectopic beats is the absence of P wave in the ECG beat morphology.  Ventricular or supraventricular beats are considered ectopic beats.  Thus, the ectopic beat detection (mostly known as incidences) is based on the regular rate of the normal beats (R peak).  Two consecutive R peaks define an RR value.  A time-varying threshold which is updated by the velocity of change of the RR time series detects those ectopic heartbeats.

Series generation

                Series and signals can represent the same.  However, a signal will be defined as a number or observations at a fixed sample rate, whereas a series would not have to.  One example could explain it.  Imaging two consecutive R peaks (which mean two consecutive heartbeats) then an RR value can be calculated as the time difference between them.  The RR value corresponds to the first heartbeat, so extending the idea one RR time series has values unevenly sampled due to the time beat occurrence.

                Series can be generated to store time events, such as RR, or to store amplitude, such as series of T wave amplitudes.   In addition, pre-processing series is also available in BioSigBrowser.

Events generation by segmentation

                An event is defined as a time interval which has associated some value.  For example, an ischemic event, which starts at some point, the percentage of ST elevation characterizes it and ends.

                The aim of segmentation is to achieve segments that accomplish the criteria that user wants.

Segmentation setup interface.

As it is shown in the Fig. 18, the criteria for segmentation that the current version of BioSigBrowser has implemented involves: outliers based on standard deviation; maximum or minimum length and memory.  Besides, two series can join the same criteria in order to both of them accomplish the requirements at the same time.

Advanced processing algorithms (Analysis)


  • Wavelet-based ECG detector/delineator: main beat waves (P, Q, R, S, T) and their onset and offset can be extracted.  Several performances are available: single lead, multi-lead based on vectorcardiogram (VCG) and multi-lead applying post-processing rules to ensure the reliability of each mark [JP2004].
  • Heart beat classifier: Normal, supraventricular and ventricular beats can be identified by this classifier. Classification can be done by three performances: fully automatic, slightly-assisted and fully-assisted [Mariano2010].
  • PPG pulse detector: analogously of detecting main beat waves, in photoplestimographic signals the detection of each pulse is the first step for further analysis [Eduardo].
  • Ischemic detector: a root mean squared over ST-T segment evaluated in the long term ST Database in three protocols. The study was driven on long recordings 21- 24 hours duration and therefore the reliability of its application on other recordings is not ensured [Ana2005].
  • T-Wave alternans detector: The quantification of T-Wave alternans is based on a multilead approach carrying out the estimation of the amplitude of the alternans [Violeta2008].
  • ECG-derived respiration (EDR):
    • VCG-based: This algorithm exploits the oscillatory pattern of the rotation angles of the heart’s electrical axis as induced by respiration. At least three orthogonal leads are needed to apply this method since it is based on the vectorcardiogram (VCG) representation [Raquel2006].
    • QRS slopes-based: This technique is based on the variations of the slopes of the QRS complexes for downward and upward slopes. Respiratory rate estimation consists in two phases: power spectrum estimation and “peak-conditioned” averaging [JesusL2013].
  • Heart rate turbulence (HRT): A Neyman-Pearson approach is used to detect and characterize HRT after ventricular premature beats.  Turbulence onset (TO) and slope (TS) are obtained from the RR intervals [JP2008].
  • Baroreflex sensitivity analysis: events and sequence techniques are available to assess the baroreflex sensitivity, nevertheless that is only possible if arterial blood pressure and ECG is acquired in the same recording [Sonia].
  • Heart rate variability analysis (HRV): Time and frequency domain for stationary signals are provided. Moreover, the time variant HRV analysis is also included by smoothed pseudo Wigner-Ville distribution [Michele].   Parametric methods such as autoregressive (AR) or ARARX are also available to HRV analysis.
  • Heart rate variability non-linear analysis: approximate and sample entropy and correlation dimension are the non-linear methods included for assessing the HRV from the chaos theory point of view [Juan].

Ventricular repolarisation adaptation to heart rate changes:  QT series (QRS onset to T wave offset) remains the ventricular repolarisation of the heart while RR means the heart rate.  An abrupt change in the heart rhythm is followed by a change in the ventricular repolarisation with a certain delay [Pueyo2004], which is related with arrhythmic risk.



Show menu

BioSigBrowser allows visualizing superimposed on the current recording, annotations and events.  Moreover, series derived from the signal can be shown in a window apart.  In addition, it exists the possibility to browse by time marks directly.

Visualizing annotations and events


                To enable the menus of visualizing annotations it is needed to load at least one annotation file, (see Loading annotation file/s).

  • On/Off: enables or disables the visualization of the annotations previously selected by user.
  • Annotation legend: each annotation file is displayed with a color code which is listed here.
  • Select annotations to show: user can select the annotations which are contained in the annotation file.
  • Change the relation between leads and annotfiles: the annotations coming from an annotation file can be superimposed on one or more leads. This setup allows changing the configuration.
  • Change visualization mode of the wave boundaries: By default the boundaries are drawn as dashed bars. Changing this visualization, closed and open parenthesis are shown instead of.


In the same way as for annotations, an file containing events has to be loaded.

  • On/Off: enables or disables the visualization of the events.
  • Change event file: The setup of “Selection of Time for Analysis Modules” will be open in order to load event file.

Visualizing series

To enable the menus of visualizing series it is needed to load at least one annotation file, (see Loading annotation file/s) containing series.

  • On/Off: enables or disables the visualization of the series.
  • Series to show: The setup of “Selection of Time for Analysis Modules” will be open in order to load event file.

Setup interface to select series contained in annotations file/s.

Next, another window will open, named Trend Window.

Trend Window showing in this case.

In the example, two series were selected.  They are displayed in two subplots.  Each series is identified by a popup menu in the middle of the graph.  This popup menu allows changing between the pre-selected series.  At the bottom a slide bar allows browsing through the series.  In addition, usually series are derived from signals, and then are linked to time positions in the signal.  Thus, the green zone that can be seen in the series corresponds to the pieze of signal that is displayed in the Signal Window.  Moreover, if user clicks on wherever in series graph automatically the Signal Window is updated at this selected time recording.

Making zoom in the series visualization is available right clicking in the graphs.  A menu appears allowing: changing the number of trends to visualize; the time scale; and an option to recover the series into the workspace of Matlab®.

Time marks navigation

This module satisfies the necessity of browsing in the recordings at specific time moments defined by marks.  For example, we used ECG detector/delineator to delineate T waves.  We know that our recording has parts with no signal due to the bad electrodes contact with skin.  So, we have in mind that T waves are not going to be consecutive and we want to check it.  We will use time marks navigation over T wave peaks to browse in the signal but in the times of the T waves.

Other example, our database has external marks given by a electrophysiologist.  The clinician marked qrs complexes as: normal (N), ventricular (V) and supraventricular (S).  And our intention is to analyze the heart rate variability. Thus, we only want normal beats (N).  If we want to know where are the other type of beats, we can use time navigation over qrs marks and then select the subtype V, for instance.  In such a way we browse in the signal just in the time positions of the V beats.

As it happens in the other visualizing modules, it is needed at least one loaded annotation file containing annotations to enable the following menus:

  • On/Off: enables or disables the time navigation mode.
  • Select mark to navigate: only annotations or events are available for time navigation.

Setup to select marks to navigate by.

  • Legends navigation: type of MIT annotations related with heartbeat classification are shown as well as rhythm change annotation labels, ST change and T change labels as well. These labels are the mostly used and are a kind of standardization.

Visualization setup for time and marks navigation.


  • Advanced options: in this setup, user can change the relative position of the mark of navigation (Left, center, right, anywhere) and also the velocity of change when auto-browse mode is on (from “as fast as”, 1 s delay to 15 s).

When time mark navigation is mode on, Signal Window shows a blue square centered in the time mark as it is shown in the Fig. 24.

BioSigBrowser when time marks navigation mode is on.

Remove marks from file

Any mark stored in an annotation file could be remove by two ways.  First, using the menu with name “Remove marks from file”.  There user will find all marks that loaded file contains.  Second, once annotations file is loaded and marks are superimposed in the signal, just right-clicking on the mark label it appears a new menu where it is possible to erase that mark or all marks belonging to that label.

Grid mode

Electrocardiogram is recorded in the clinical practice historically printed in paper.  Grid representation allows clinicians to measure in-situ amplitude as well as time intervals.  This mode is in such a way the classical representation of electrocardiography recording.

Signal Info menu

One single recording consists in several features that only signal representation cannot reflect.  Information of sampling frequency, number of leads/signals, number of samples, and so on, is shown in the menu “General Info”.

While user works with BioSigBrowser interface to test some filters, for instance, the signal shown in the Signal Window changes with respect to the original one.  In the menu “Current Signal Info” are described a brief description of the pre-processing methods applied to the signals.  If user wants to remove all pre-processing steps done till that moment, just going to pre-processing menu and select “Go to original signal”.

In addition, some database formats allow patient description.  In this case, the related information is available in the menu “Recorder comments”.

Batchmode Assistant

BioSigBrowser is a graphical interface that allows applying several methodologies in a friendly framework.  However, most of the algorithms can be used out of the interface and they can be anchored generating specific workflows.  This tool is very useful for automatizating processes.  For example, to delineate a whole database.

Batchmode assistant is also a friendly way to construct workflows.  Nevertheless, in the current version of BioSigBrowser user may need some help to make it works properly.  Do not hesitate to ask for help in the following e-mail account:

The Batchmode Assistant is divided into steps that allows user to construct a workflow well defined.  Let see, the available modules.

  • Step 1/8: The workflow is going to be written in an XML file.


Setup of step 1/8 of Batchmode Assistant.


  • Step 2/8: This step is thought to produce signals from series and/or annotations.  For example, user did EDR based on QRS slopes, so, the respiratory rate as series is available.  It could be used to reconstruct the signal of respiration.


Setup of step 2/8 of Batchmode Assistant.


  • Step 3/8: Signal pre-processing step.  After applying pre-processing modules, the new signal should be save in order to the following methods will be applied to the new one.

Setup of step 3/8 of Batchmode Assistant.


  • Step 4/8: This step consists in producing annotations from signal and/or series.  All kind of detectors such as ECG detector, PPG artifact or pulse detector are examples.


Setup of step 4/8 of Batchmode Assistant.


  • Step 5/8: This module allows pre-processing annotations.

Setup of step 5/8 of Batchmode Assistant.


  • Step 6/8: This step is based on producing series coming through annotations and/or signals.


Setup of step 6/8 of Batchmode Assistant.


  • Step 7/8: Pre-processing of series in the terms of detrending or outlier treatment are the methods available in this module.

Setup of step 7/8 of Batchmode Assistant.


  • Step 8/8: Once user has defined the pre-processing steps to adecuate the framework, the advance processing algorithms that are shown in Fig. X are available in batchmode.

Setup of step 8/8 of Batchmode Assistant.

Los comentarios están cerrados