Nitin Burman
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Nitin Burman has a total of 7 co-authors. Here are the top co-authors by number of shared publications:
- Jan D'Hooge
Department of Cardiovascular sciences, KU Leuven, Leuven, Belgium
2 shared publications - Andrea Pulido
Department of Cardiovascular sciences, KU Leuven, Leuven, Belgium
1 shared publication - Claudia Alessandra Manetti
Department of Biomedical Engineering, CARIM School for Cardiovascular Diseases, Maastricht University Medical Center , Universiteitssingel 40, 6229 ERMaastricht
1 shared publication - Joost Lumens
Department of Biomedical Engineering, Cardiovascular Research Institute Maastricht (CARIM), Maastricht University , Maastricht
1 shared publication - Marcus Ingram
Department of Cardiovascular sciences, KU Leuven, Leuven, Belgium
1 shared publication - Sandro Queirós
Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal
1 shared publication
Latest Publications
- Open Access
- 5 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of Ultrasonics, Ferroelectrics, and Frequency Control
- Vol. 5, 286-291
Impact of Training Data Composition on Deep Learning-Based Cardiac Motion Estimation
Cardiac ultrasound deformation imaging is a non-invasive and widely used modality for assessing left ventricular function. However, due to the variable image quality observed in clinical routine, deformation estimation remains challenging. Recently, deep learning (DL) approaches have been proposed to estimate motion from ultrasound images. In this study, we investigated whether combining in vivo data with a synthetic dataset during training improves motion estimation accuracy while also benchmarking the DL-based estimates against a state-of-the-art traditional optical flow method. Results demonstrate that the deep learning-based method excels in tracking the dense motion fields on a frame-by-frame basis. However, when performing contour tracking, it results in lower end-systolic peak strain values compared to the traditional optical flow method.
- 5 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of Ultrasonics, Ferroelectrics, and Frequency Control
- Vol. 5, 286-291
Open AccessImpact of Training Data Composition on Deep Learning-Based Cardiac Motion Estimation
Cardiac ultrasound deformation imaging is a non-invasive and widely used modality for assessing left ventricular function. However, due to the variable image quality observed in clinical routine, deformation estimation remains challenging. Recently, deep learning (DL) approaches have been proposed to estimate motion from ultrasound images. In this study, we investigated whether combining in vivo data with a synthetic dataset during training improves motion estimation accuracy while also benchmarking the DL-based estimates against a state-of-the-art traditional optical flow method. Results demonstrate that the deep learning-based method excels in tracking the dense motion fields on a frame-by-frame basis. However, when performing contour tracking, it results in lower end-systolic peak strain values compared to the traditional optical flow method.
- Open Access
- 21 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 117040-117055
Large-Scale Simulation of Realistic Cardiac Ultrasound Data With Clinical Appearance: Methodology and Open-Access Database
Cardiac ultrasound imaging is widely used in the clinical setting. Deep learning algorithms have shown increased potential in automating routine clinical tasks for improved diagnosis and prognosis. However, the lack of large and diverse training databases hinders their employment on a full scale in clinics. Difficulties associated with creating large training databases include scarcity of clinical data, privacy related data access concerns and labor-intensive ground-truth annotations. To overcome these issues, realistic cardiac ultrasound simulation is considered advantageous. Nevertheless, our state-of-the-art cardiac ultrasound simulation pipeline is limited to small-scale databases particularly because of the underlying computationally complex finite-element heart model. In this paper, we developed a realistic simulation pipeline to facilitate large-scale synthetic cardiac ultrasound datasets. This pipeline utilizes a well-validated and computationally efficient lump-parameter heart model for generating cardiovascular mechanics, and the convolution based ultrasound simulator COLE for short simulation times. We implemented the proposed pipeline to create a large open-access database of 2-D realistic cardiac ultrasound recordings with diversity in the geometries, ultrasound speckle texture and motions of the left ventricle. Quantitative assessment of the simulated data demonstrated that the synthetic cardiac ultrasound frames contained realistic speckle patterns. Moreover, the left ventricular motion in the simulated frames was tracked using the medical imaging tracking toolbox and was shown to closely match the kinematically imposed ground-truth left ventricular motion. The presented pipeline is valuable for creating large and diverse ground-truth training databases. The resulting database is a potential augmentation tool for machine learning-based ultrasound data processing algorithms and made publicly available.
- 21 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 117040-117055
Open AccessLarge-Scale Simulation of Realistic Cardiac Ultrasound Data With Clinical Appearance: Methodology and Open-Access Database
Cardiac ultrasound imaging is widely used in the clinical setting. Deep learning algorithms have shown increased potential in automating routine clinical tasks for improved diagnosis and prognosis. However, the lack of large and diverse training databases hinders their employment on a full scale in clinics. Difficulties associated with creating large training databases include scarcity of clinical data, privacy related data access concerns and labor-intensive ground-truth annotations. To overcome these issues, realistic cardiac ultrasound simulation is considered advantageous. Nevertheless, our state-of-the-art cardiac ultrasound simulation pipeline is limited to small-scale databases particularly because of the underlying computationally complex finite-element heart model. In this paper, we developed a realistic simulation pipeline to facilitate large-scale synthetic cardiac ultrasound datasets. This pipeline utilizes a well-validated and computationally efficient lump-parameter heart model for generating cardiovascular mechanics, and the convolution based ultrasound simulator COLE for short simulation times. We implemented the proposed pipeline to create a large open-access database of 2-D realistic cardiac ultrasound recordings with diversity in the geometries, ultrasound speckle texture and motions of the left ventricle. Quantitative assessment of the simulated data demonstrated that the synthetic cardiac ultrasound frames contained realistic speckle patterns. Moreover, the left ventricular motion in the simulated frames was tracked using the medical imaging tracking toolbox and was shown to closely match the kinematically imposed ground-truth left ventricular motion. The presented pipeline is valuable for creating large and diverse ground-truth training databases. The resulting database is a potential augmentation tool for machine learning-based ultrasound data processing algorithms and made publicly available.