Robust Cardiac Motion Estimation With Dictionary Learning and Temporal Regularization for Ultrasound Imaging

Ouzir, Nora; Bioucas-Dias, José M.; Basarab, Adrian; Tourneret, Jean-Yves

doi:10.1109/ultsym.2019.8925936

Cited by 4 publications

(5 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We hypothesize that not incorporating these complex imaging conditions in our simulation framework might resulted in the performance difference observed between simulation and in vivo . Furthermore, the better performance in vivo due to temporal coherence corroborates with findings from previous literature reports [ 9 ]–[ 11 ], [ 13 ], [ 30 ].…”

Section: Discussionsupporting

confidence: 90%

“…In this work, temporal consistency is designed to be piecewise smooth as information from only immediate past and future neighboring frames is used in contrast to spatiotemporal algorithms which enforce global smooth displacement trajectory over a cardiac cycle [9] with cyclic periodicity [71]. Similar, piecewise temporal smoothness was previously reported in literature for CSI [11,13]. This constraint is applied to ensure robustness against out-of-plane motion artifacts which can introduce large discontinuity in temporal displacement fields consequently impacting the cardiac strain estimates negatively thus resulting in physiologically more plausible strain variation in vivo [observe Figure 12 and 14].…”

Section: Discussionmentioning

confidence: 91%

“…al . proposed spatial and sparse regularization with dictionary learning and reported better motion estimation accuracy compared to state-of-the-art methods [ 12 ], which was extended to incorporate temporal information [ 13 ]. Despite regularization being inherently embedded in these NRIR-based methods, they suffer from reduced sensitivity to small inter-frame displacements and lower elastographic signal-to-noise ratio (SNR e ) due to the use of US B-mode or envelope data instead of RF data [ 14 ].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Spatiotemporal Bayesian Regularization for Cardiac Strain Imaging: Simulation and In Vivo Results

Mukaddim¹,

Meshram

Weichmann

et al. 2021

IEEE Open J. Ultrason., Ferroelect., Freq. Contr.

View full text Add to dashboard Cite

Cardiac strain imaging (CSI) plays a critical role in the detection of myocardial motion abnormalities. Displacement estimation is an important processing step to ensure the accuracy and precision of derived strain tensors. In this paper, we propose and implement Spatiotemporal Bayesian regularization (STBR) algorithms for two-dimensional (2-D) normalized cross-correlation (NCC) based multi-level block matching along with incorporation into a Lagrangian cardiac strain estimation framework. Assuming smooth temporal variation over a short span of time, the proposed STBR algorithm performs displacement estimation using at least four consecutive ultrasound radio-frequency (RF) frames by iteratively regularizing 2-D NCC matrices using information from a local spatiotemporal neighborhood in a Bayesian sense. Two STBR schemes are proposed to construct Bayesian likelihood functions termed as Spatial then Temporal Bayesian (STBR-1) and simultaneous Spatiotemporal Bayesian (STBR-2). Radial and longitudinal strain estimated from a finite-element-analysis (FEA) model of realistic canine myocardial deformation were utilized to quantify strain bias, normalized strain error and total temporal relative error (TTR). Statistical analysis with one-way analysis of variance (ANOVA) showed that all Bayesian regularization methods significantly outperform NCC with lower bias and errors ( p < 0.001 ). However, there was no significant difference among Bayesian methods. For example, mean longitudinal TTR for NCC, SBR, STBR-1 and STBR-2 were 25.41%, 9.27%, 10.38% and 10.13% respectively An in vivo feasibility study using RF data from ten healthy mice hearts were used to compare the elastographic signal-to-noise ratio (SNR e ) calculated using stochastic analysis. STBR-2 had the highest expected SNR e both for radial and longitudinal strain. The mean expected SNR e values for accumulated radial strain for NCC, SBR, STBR-1 and STBR-2 were 5.03, 9.43, 9.42 and 10.58, respectively. Overall results suggest that STBR improves CSI in vivo .

show abstract

Section: Discussionsupporting

confidence: 90%

Section: Discussionmentioning

confidence: 91%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Spatiotemporal Bayesian Regularization for Cardiac Strain Imaging: Simulation and In Vivo Results

Mukaddim¹,

Meshram

Weichmann

et al. 2021

IEEE Open J. Ultrason., Ferroelect., Freq. Contr.

View full text Add to dashboard Cite

show abstract

“…The majority of techniques employed for this task primarily rely on imaging modalities. These may include: fluoroscopy [49], ultrasound [50], computed tomography (CT) [51], camera vision [52], or ECG-gated magnetic resonance (MR) [53]. While the underlying techniques and algorithms may differ between the aforementioned imaging modalities, the general work-flow is similar.…”

Section: A Local Heart Motion Estimationmentioning

confidence: 99%

Force Control With a Novel Robotic Catheterization System Based on Braided Sleeve Grippers

Al-Ahmad

Ourak

Vlekken³

et al. 2023

IEEE Trans. Med. Robot. Bionics

View full text Add to dashboard Cite

The complex, deformable, and dynamic cardiovascular system makes precise control of flexible medical instruments a challenging task. An innovative robot-assisted catheterization system (RACS) based on braided sleeves was thus developed to aid interventionalists during cardiovascular procedures. The RACS allows navigating instruments through the vasculature with continuous uninterrupted motion. The complete design and characterization of the developed RACS are presented and experimentally evaluated. The RACS' capability to track dynamic motion is demonstrated. Hereto, heart motion is estimated and fed into a dedicated motion controller. It is further demonstrated that tissue tracking performance can be improved by compensating for hysteresis present along the instrument's length and incorporating a rate-dependent Bouc-Wen model in the motion control strategy. Here, a practical method based on multi-core FBG-inscribed fiber technology is presented for continuous instrument pose tracking. A velocity controller is cascaded with a force control loop to enable instrument tip contact force tracking. An experimental comparison between three different force control strategies was carried out on a benchtop laboratory setup resembling a dynamically moving heart wall. Results confirmed the superior performance of the cascaded force-velocity controller with Bouc-Wen hysteresis compensation. Finally, the RACS system was validated through in-vivo experiments on a living swine which demonstrated its ability to successfully navigate instruments toward the heart.

show abstract

“…The topic of heart motion estimation has already been tackled in several prior works. In general, the majority of techniques employed for this task rely on one the following imaging modalities: a) fluoroscopy [9], [10], b) ultrasound [11], [12], c) computed tomography (CT) [13], [14], d) camera vision [15], [16], or e) ECG-gated magnetic resonance (MR) [17]- [19]. While the underlying techniques and algorithms may differ between the aforementioned imaging modalities, the general work-flow is similar.…”

Section: B Prior Workmentioning

confidence: 99%

Local One-Dimensional Motion Estimation Using FBG-Based Shape Sensing for Cardiac Applications

Al-Ahmad

Ourak

Vlekken³

et al. 2022

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

The human heart is a fragile and dynamic organ that requires careful approach during catheter intervention. It is essential for physicians to have a high degree of awareness of the anatomy and the relative pose of the catheter when operating inside the heart. One of the key aspects that can aid physicians during such interventions is knowledge of the heart's motion profile. Accordingly, this work addresses the objective of estimating local predominantly one-dimensional heart motions by making use of FBG-inscribed multi-core fibers and shape sensing. An FBG-fiber embedded in a cardiac catheter is propagated towards a designated region in the heart, where a force sensor at the tip of the catheter is used to identify contact with the heart tissue. The catheter's tip position is continuously monitored during contact with the moving tissue. The proposed method is able to determine the direction of the local surface motion and estimates the evolution of the motion profile through time. An Unscented Kalman Filter (UKF) is thus employed to provide continuous quasi-periodic estimation of this motion. Experiments on a bench-top laboratory heart mock-up were carried out to validate the proposed approach. Results show that the proposed method can provide accurate estimations of the heart motion profile with an absolute mean error of 1.1 ± 0.4 mm (9.2%) for motions with average peak-to-peak amplitudes of 12 mm.

show abstract

Robust Cardiac Motion Estimation With Dictionary Learning and Temporal Regularization for Ultrasound Imaging

Abstract: Robust cardiac motion estimation with dictionary learning and temporal regularization for ultrasound imaging.

Cited by 4 publications

References 15 publications

Spatiotemporal Bayesian Regularization for Cardiac Strain Imaging: Simulation and In Vivo Results

Spatiotemporal Bayesian Regularization for Cardiac Strain Imaging: Simulation and In Vivo Results

Force Control With a Novel Robotic Catheterization System Based on Braided Sleeve Grippers

Local One-Dimensional Motion Estimation Using FBG-Based Shape Sensing for Cardiac Applications

Contact Info

Product

Resources

About