Related Experiment Video
Updated: Nov 22, 2025
Quantification of Mouse Heart Left Ventricular Function, Myocardial Strain, and Hemodynamic Forces by Cardiovascular Magnetic Resonance Imaging
Published on: May 24, 2021
Deep Layer Kernel Sparse Representation Network for the Detection of Heart Valve Ailments from the Time-Frequency
Samit Kumar Ghosh1, R N Ponnalagu1, R K Tripathy1
1Department of Electrical and Electronics Engineering, BITS-Pilani, Hyderabad Campus, Hyderabad 500078, India.
Insights
Automated early detection of heart valve ailments (HVAs) is crucial for timely treatment and reduced mortality. A novel deep learning network accurately classifies various HVAs from phonocardiogram (PCG) signals, achieving high sensitivity for normal and pathological cases.
Area of Science:
- Cardiology
- Biomedical Engineering
- Artificial Intelligence
Background:
- Heart valve ailments (HVAs) can lead to severe complications including heart failure and sudden cardiac death if left untreated.
- Early and accurate detection of HVAs is essential for effective patient management and improved outcomes.
- Phonocardiogram (PCG) signals capture heart sounds and murmurs, offering a non-invasive method for assessing heart valve function.
Purpose of the Study:
- To propose a novel time-frequency-based deep layer kernel sparse representation network (DLKSRN) for automated detection of various HVAs.
- To evaluate the efficacy of the DLKSRN in classifying normal and pathological heart sound recordings.
- To assess the potential of the DLKSRN for integration into Internet of Things- (IoT-) driven smart healthcare systems.
Main Methods:
- Utilized Spline kernel-based Chirplet Transform (SCT) for time-frequency representation of PCG signals.
- Extracted features including L1-norm (LN), sample entropy (SEN), and permutation entropy (PEN) from the time-frequency representations.
- Employed a DLKSRN, integrating extreme learning machine (ELM) autoencoders and kernel sparse representation (KSR), for classifying PCG recordings.
Main Results:
- The DLKSRN achieved high average sensitivities: 100% for normal cases, 97.51% for mitral valve prolapse (MVP), 99.00% for mitral regurgitation (MR), 98.72% for aortic stenosis (AS), and 99.13% for mitral stenosis (MS).
- The method demonstrated robust performance across both public and private PCG databases.
- Hold-out cross-validation (CV) was employed for performance evaluation.
Conclusions:
- The proposed DLKSRN offers an accurate and automated approach for detecting heart valve ailments using PCG signals.
- The method shows significant potential for real-time HVA detection in smart healthcare applications.
- This automated system can aid clinicians in timely diagnosis and treatment, potentially reducing mortality rates associated with HVAs.
Abstract:
The heart valve ailments (HVAs) are due to the defects in the valves of the heart and if untreated may cause heart failure, clots, and even sudden cardiac death. Automated early detection of HVAs is necessary in the hospitals for proper diagnosis of pathological cases, to provide timely treatment, and to reduce the mortality rate. The heart valve abnormalities will alter the heart sound and murmurs which can be faithfully captured by phonocardiogram (PCG) recordings. In this paper, a time-frequency based deep layer kernel sparse representation network (DLKSRN) is proposed for the detection of various HVAs using PCG signals. Spline kernel-based Chirplet transform (SCT) is used to evaluate the time-frequency representation of PCG recording, and the features like L1-norm (LN), sample entropy (SEN), and permutation entropy (PEN) are extracted from the different frequency components of the time-frequency representation of PCG recording. The DLKSRN formulated using the hidden layers of extreme learning machine- (ELM-) autoencoders and kernel sparse representation (KSR) is used for the classification of PCG recordings as normal, and pathology cases such as mitral valve prolapse (MVP), mitral regurgitation (MR), aortic stenosis (AS), and mitral stenosis (MS). The proposed approach has been evaluated using PCG recordings from both public and private databases, and the results demonstrated that an average sensitivity of 100%, 97.51%, 99.00%, 98.72%, and 99.13% are obtained for normal, MVP, MR, AS, and MS cases using the hold-out cross-validation (CV) method. The proposed approach is applicable for the Internet of Things- (IoT-) driven smart healthcare system for the accurate detection of HVAs.
