Related Experiment Video
Updated: Oct 29, 2025
Direct Comparison of Hyperspectral Stimulated Raman Scattering and Coherent Anti-Stokes Raman Scattering Microscopy for Chemical Imaging
Published on: April 28, 2022
A comprehensive comparison between shortest-path HARP refinement, SinMod, and DENSEanalysis processing tools applied
Hernán Mella1, Joaquín Mura2, Julio Sotelo3
1Department of Electrical Engineering, Pontificia Universidad Católica de Chile, Santiago, Chile; Biomedical Imaging Centre, Pontificia Universidad Católica de Chile, Santiago, Chile; Millennium Nucleus for Cardiovascular Magnetic Resonance, Santiago, Chile.
DENSEanalysis demonstrated superior accuracy in estimating cardiac strain compared to SP-HR and SinMod, especially under challenging conditions like noise and motion artifacts in cardiac MRI.
Area of Science:
- Biomedical Engineering
- Medical Imaging
- Cardiovascular Mechanics
Background:
- Accurate assessment of myocardial strain is crucial for diagnosing cardiac conditions.
- Existing motion estimation techniques face challenges with image artifacts, noise, and complex tissue motion.
Purpose of the Study:
- To comprehensively evaluate and compare the performance of three cardiac motion estimation techniques: Shortest-Path HARP Refinement (SP-HR), SinMod, and DENSEanalysis.
- To assess their accuracy under various imaging conditions, including noise, artifacts, and through-plane motion.
Main Methods:
- Utilized 2D slices of synthetic CSPAMM and DENSE cardiac magnetic resonance (MR) images derived from 3D phantoms.
- Tested SP-HR, SinMod, and DENSEanalysis under ideal and non-ideal conditions (noise, MR artifacts, through-plane motion) at multiple resolutions and noise levels.
- Quantified performance using normalized root-mean-square error (nRMSE) for circumferential and radial strain components.
Main Results:
- DENSEanalysis achieved the lowest nRMSE for both circumferential (6.4 ± 2.0%) and radial (18.2 ± 4.6%) strain under noisy conditions.
- SP-HR showed limitations with large tissue motions.
- SinMod and DENSEanalysis provided accurate displacement and strain field estimations, with DENSEanalysis outperforming SinMod.
Conclusions:
- DENSEanalysis is the most accurate technique for myocardial strain estimation among the evaluated methods, particularly in the presence of noise and motion.
- SP-HR is less reliable for significant tissue motion.
- SinMod and DENSEanalysis offer robust performance for cardiac motion analysis.
