Digitala Vetenskapliga Arkivet

Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Assessment of Valvular Aortic Stenosis by Signal Analysis of the Phonocardiogram
Linköping University, Department of Biomedical Engineering. Linköping University, The Institute of Technology.
2014 (English)Doctoral thesis, comprehensive summary (Other academic)
Abstract [en]

Aortic stenosis (AS) is one of the most prevalent valvular heart diseases in elderly people. According to the recommendations of both the American Heart Association and the European Society of Cardiology, severity assessment of AS is primarily based on echocardiographic findings. The experience of the investigator here play important roles in the accuracy of the assessment, and therefore in the disease management. However, access to the expert physicians could be limited, especially in rural health care centers of developing countries.

This thesis aims to develop processing algorithms tailored for phonocardiographic signal with the intension to obtain a noninvasive diagnostic tool for AS assessment and severity grading. The algorithms employ a phonocardiogram as input signal and perform analysis for screening and diagnostics. Such a decision support system, which we call “the intelligent phonocardiography”, can be widely used in primary healthcare centers.

The main contribution of the thesis is to present innovative models for the phonocardiographic analysis by taking the segmental characteristics of the signal into consideration. Three novel methodologies are described, based on the presented models, to perform robust classification. In the first attempt, a novel pattern recognition framework is presented for screening of AS-related murmurs. The framework offers a hybrid model for classifying cyclic time series in general, but is tailored to detect the murmurs as a special case study. The time growing neural network is another method that we use to classify short time signals with abrupt frequency transition. The idea of the growing frames is extended to the cyclic signals with stochastic properties for the screening purposes. Finally, a combined statistical and artificial intelligent classifier is proposed for grading the severity of AS.

The study suggests comprehensive statistical validations not only for the evaluation and representation of systolic murmurs but also for setting the methodology design parameters, which can be considered as one of the significant features of the study. The resulting methodologies can be implemented by using web and mobile technologies to be utilized in distributed healthcare system.

Place, publisher, year, edition, pages
Linköping: Linköping University Electronic Press, 2014. , p. 81
Series
Linköping Studies in Science and Technology. Dissertations, ISSN 0345-7524 ; 1616
National Category
Medical Biotechnology Medical Bioscience
Identifiers
URN: urn:nbn:se:liu:diva-110182ISBN: 978-91-7519-252-9 (print)OAI: oai:DiVA.org:liu-110182DiVA, id: diva2:743395
Public defence
2014-09-26, Linden, Campus US, Linköpings universitet, Linköping, 10:00 (English)
Opponent
Supervisors
Available from: 2014-09-04 Created: 2014-09-04 Last updated: 2016-12-28Bibliographically approved
List of papers
1. A pattern recognition framework for detecting dynamic changes on cyclic time series
Open this publication in new window or tab >>A pattern recognition framework for detecting dynamic changes on cyclic time series
2015 (English)In: Pattern Recognition, ISSN 0031-3203, E-ISSN 1873-5142, Vol. 48, no 3, p. 696-708Article in journal (Refereed) Published
Abstract [en]

This paper proposes a framework for binary classification of the time series with cyclic characteristics. The framework presents an iterative algorithm for learning the cyclic characteristics by introducing the discriminative frequency bands (DFBs) using the discriminant analysis along with k-means clustering method. The DFBs are employed by a hybrid model for learning dynamic characteristics of the time series within the cycles, using statistical and structural machine learning techniques. The framework offers a systematic procedure for finding the optimal design parameters associated with the hybrid model. The proposed  model is optimized to detect the changes of the heart sound recordings (HSRs) related to aortic stenosis. Experimental results show that the proposed framework provides efficient tools for classification of the HSRs based on the heart murmurs. It is also evidenced that the hybrid model, proposed by the framework, substantially improves the classification performance when it comes to detection of the heart disease.

Place, publisher, year, edition, pages
Elsevier, 2015
Keywords
Hybrid model, cyclic time series, time series, phonocardiogram, systolic murmurs
National Category
Biomedical Laboratory Science/Technology Medical Biotechnology
Identifiers
urn:nbn:se:liu:diva-110177 (URN)10.1016/j.patcog.2014.08.017 (DOI)000347747000008 ()
Available from: 2014-09-04 Created: 2014-09-04 Last updated: 2017-12-05Bibliographically approved
2. Detection of systolic ejection click using time growing neural network
Open this publication in new window or tab >>Detection of systolic ejection click using time growing neural network
2014 (English)In: Medical Engineering and Physics, ISSN 1350-4533, E-ISSN 1873-4030, Vol. 36, no 4, p. 477-483Article in journal (Refereed) Published
Abstract [en]

In this paper, we present a novel neural network for classification of short-duration heart sounds: the time growing neural network (TGNN). The input to the network is the spectral power in adjacent frequency bands as computed in time windows of growing length. Children with heart systolic ejection click (SEC) and normal children are the two groups subjected to analysis. The performance of the TGNN is compared to that of a time delay neural network (TDNN) and a multi-layer perceptron (MLP), using training and test datasets of similar sizes with a total of 614 normal and abnormal cardiac cycles. From the test dataset, the classification rate/sensitivity is found to be 97.0%/98.1% for the TGNN, 85.1%/76.4% for the TDNN, and 92.7%/85.7% for the MLP. The results show that the TGNN performs better than do TDNN and MLP when frequency band power is used as classifier input. The performance of TGNN is also found to exhibit better immunity to noise.

Place, publisher, year, edition, pages
Elsevier, 2014
Keywords
Systolic ejection click; Time growing neural network; Time delay neural network; Heart sound
National Category
Engineering and Technology
Identifiers
urn:nbn:se:liu:diva-106865 (URN)10.1016/j.medengphy.2014.02.011 (DOI)000334976800008 ()
Available from: 2014-05-28 Created: 2014-05-23 Last updated: 2017-12-05
3. A novel method for discrimination between innocent and pathological heart murmurs
Open this publication in new window or tab >>A novel method for discrimination between innocent and pathological heart murmurs
2015 (English)In: Medical Engineering and Physics, ISSN 1350-4533, E-ISSN 1873-4030, Vol. 37, no 7, p. 674-682Article in journal (Refereed) Published
Abstract [en]

This paper presents a novel method for discrimination between innocent and pathological murmurs using the growing time support vector machine (GTSVM). The proposed method is tailored for characterizing innocent murmurs (IM) by putting more emphasis on the early parts of the signal as IMs are often heard in early systolic phase. Individuals with mild to severe aortic stenosis (AS) and IM are the two groups subjected to analysis, taking the normal individuals with no murmur (NM) as the control group. The AS is selected due to the similarity of its murmur to IM, particularly in mild cases. To investigate the effect of the growing time windows, the performance of the GTSVM is compared to that of a conventional support vector machine (SVM), using repeated random sub-sampling method. The mean value of the classification rate/sensitivity is found to be 88%/86% for the GTSVM and 84%/83% for the SVM. The statistical evaluations show that the GTSVM significantly improves performance of the classification as compared to the SVM.

Place, publisher, year, edition, pages
Elsevier, 2015
Keywords
Growing-time support vector machine, support vector machine, phonocardiogram signal, heart murmurs, innocent murmurs.
National Category
Medical Engineering
Identifiers
urn:nbn:se:liu:diva-117825 (URN)10.1016/j.medengphy.2015.04.013 (DOI)000357354400007 ()26003286 (PubMedID)
Note

At the time for thesis presentation publication was in status: Manuscript

Available from: 2015-05-08 Created: 2015-05-08 Last updated: 2017-12-04Bibliographically approved
4. An Automatic Tool for Pediatric Heart Sounds Segmentation
Open this publication in new window or tab >>An Automatic Tool for Pediatric Heart Sounds Segmentation
Show others...
(English)Manuscript (preprint) (Other academic)
Abstract [en]

In this paper, we present a novel algorithm for pediatric heart sound segmentation, incorporated into a graphical user interface. The algorithm employs both the Electrocardiogram (ECG) and Phonocardiogram (PCG) signals for an efficient segmentation under pathological circumstances.First, the ECG signal is invoked in order to determine the beginning and end points of each cardiac cycle by using wavelet transform technique. Then, first and second heart sounds within the cycles are identified over the PCG signal by paying attention to the spectral properties of the sounds. The algorithm is applied on 120 recordings of normal and pathological children, totally containing 1976 cardiac cycles. The accuracy of the segmentation algorithm is 97% for S1 and 94% for S2 identification while all the cardiac cycles are correctly determined.

National Category
Biomedical Laboratory Science/Technology Medical Biotechnology
Identifiers
urn:nbn:se:liu:diva-110179 (URN)
Available from: 2014-09-04 Created: 2014-09-04 Last updated: 2014-09-04Bibliographically approved
5. Severity assessments of aortic stenosis using intelligent phonocardiography
Open this publication in new window or tab >>Severity assessments of aortic stenosis using intelligent phonocardiography
Show others...
(English)Manuscript (preprint) (Other academic)
Abstract [en]

Objectives: To study capabilities of the intelligent phonocardiography (IPCG) in automatic grading severity of the aortic stenosis (AS).

Methods: Phonocardiogram signals were recorded from the patients with AS, as diagnosed by echocardiography. The patient group is comprised of signals, recorded from 5 patients (2 recordings from each), mostly elderly referrals (>60 years) with mild to severe AS. An advanced processing algorithm, consisted of the wavelet transform and the stepwise regression analysis, characterizes the systolic murmur caused by the AS in order to predict the 5 indicators; mean pressure gradient over the aortic valve (MPG), maximum jet velocity (MJV), aortic valve area (AVA), velocity time integral and the ejection period. The automatic assessment is performed by an artificial neural network using the predicted values of the indicators as the input data. Reliability of the IPCG is validated by applying repeated random sub-sampling (RRSS) with 70%/30% of the training/testing data, and calculating the accuracy. The RRSS is also employed to validate reproducibility of the IPCG by using 70% of the signals for training and the second recording of the same individuals for  testing.

Results: Accuracy of the IPCG is estimated to be and (95% confidence interval) for the reliability and the reproducibility, respectively. Linear correlation between the characterized systolic murmur and the MPG (r>0.81), the MJV (r>0.82) and the AVA (r>0.85) is observed.

Conclusions: The IPCG has the potential to objectively serve as a clinical tool for grading severity of the aortic stenosis.

National Category
Biomedical Laboratory Science/Technology Medical Biotechnology
Identifiers
urn:nbn:se:liu:diva-110181 (URN)
Available from: 2014-09-04 Created: 2014-09-04 Last updated: 2014-09-04Bibliographically approved

Open Access in DiVA

omslag(137 kB)124 downloads
File information
File name COVER01.pdfFile size 137 kBChecksum SHA-512
343f3f2d114dca7e9fea5b30056ad6c48d55a1a7f4ac8f7f5bb753156365ac72e9fe83da0e7bc7dd9169dcc6d0bf379566ca36b16a11282daa8247ce7ea29a80
Type coverMimetype application/pdf

Search in DiVA

By author/editor
Ghareh Baghi, Ghareh Baghi
By organisation
Department of Biomedical EngineeringThe Institute of Technology
Medical BiotechnologyMedical Bioscience

Search outside of DiVA

GoogleGoogle Scholar
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

isbn
urn-nbn

Altmetric score

isbn
urn-nbn
Total: 1771 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf