More Like this

1. Automated Segmentation and 4-D reconstruction of the Heart Left Ventricle from CINE MRI

   Molina, Giovanni; Velazco-Garcia, Jose D.; Shah, Dipan; Becker, Aaron T.; Tsekos, Nikolaos V. (October 2019, IEEE International Conference on Bioinformatics and Bioengineering (BIBE))

   Heart disease is highly prevalent in developed countries, causing 1 in 4 deaths. In this work we propose a method for a fully automated 4D reconstruction of the left ventricle of the heart. This can provide accurate information regarding the heart wall motion and in particular the hemodynamics of the ventricles. Such metrics are crucial for detecting heart function anomalies that can be an indication of heart disease. Our approach is fast, modular and extensible. In our testing, we found that generating the 4D reconstruction from a set of 250 MRI images takes less than a minute. The amount of time saved as a result of our work could greatly benefit physicians and cardiologist as they diagnose and treat patients. Index Terms—Magnetic Resonance Imaging, segmentation, reconstruction, cardiac, machine learning, ventricle 

   more » « less
2. DeepOrganNet: On-the-Fly Reconstruction and Visualization of 3D / 4D Lung Models from Single-View Projections by Deep Deformation Network

   Wang, Yifan; Zhong, Zichun; Hua, Jing (January 2020, IEEE transactions on visualization and computer graphics)

   This paper introduces a deep neural network based method, i.e., DeepOrganNet, to generate and visualize fully high-fidelity 3D / 4D organ geometric models from single-view medical images with complicated background in real time. Traditional 3D / 4D medical image reconstruction requires near hundreds of projections, which cost insufferable computational time and deliver undesirable high imaging / radiation dose to human subjects. Moreover, it always needs further notorious processes to segment or extract the accurate 3D organ models subsequently. The computational time and imaging dose can be reduced by decreasing the number of projections, but the reconstructed image quality is degraded accordingly. To our knowledge, there is no method directly and explicitly reconstructing multiple 3D organ meshes from a single 2D medical grayscale image on the fly. Given single-view 2D medical images, e.g., 3D / 4D-CT projections or X-ray images, our end-to-end DeepOrganNet framework can efficiently and effectively reconstruct 3D / 4D lung models with a variety of geometric shapes by learning the smooth deformation fields from multiple templates based on a trivariate tensor-product deformation technique, leveraging an informative latent descriptor extracted from input 2D images. The proposed method can guarantee to generate high-quality and high-fidelity manifold meshes for 3D / 4D lung models; while, all current deep learning based approaches on the shape reconstruction from a single image cannot. The major contributions of this work are to accurately reconstruct the 3D organ shapes from 2D single-view projection, significantly improve the procedure time to allow on-the-fly visualization, and dramatically reduce the imaging dose for human subjects. Experimental results are evaluated and compared with the traditional reconstruction method and the state-of-the-art in deep learning, by using extensive 3D and 4D examples, including both synthetic phantom and real patient datasets. The efficiency of the proposed method shows that it only needs several milliseconds to generate organ meshes with 10K vertices, which has great potential to be used in real-time image guided radiation therapy (IGRT). 

   more » « less
3. Experimental studies of spiral wave teleportation in a light sensitive Belousov–Zhabotinsky system

   https://doi.org/10.1063/5.0216649  

   Tyler, Shannyn A; Mersing, David; Fenton, Flavio H; Tinsley, Mark R; Showalter, Kenneth (September 2024, Chaos: An Interdisciplinary Journal of Nonlinear Science)

   Cardiac arrythmias are a form of heart disease that contributes toward making heart disease a significant cause of death globally. Irregular rhythms associated with cardiac arrythmias are thought to arise due to singularities in the heart tissue that generate reentrant waves in the underlying excitable medium. A normal approach to removing such singularities is to apply a high voltage electric shock, which effectively resets the phase of the cardiac cells. A concern with the use of this defibrillation technique is that the high-energy shock can cause lasting damage to the heart tissue. Various theoretical works have investigated lower-energy alternatives to defibrillation. In this work, we demonstrate the effectiveness of a low-energy defibrillation method in an experimental 2D Belousov–Zhabotinsky (BZ) system. When implemented as a 2D spatial reaction, the BZ reaction serves as an effective analog of general excitable media and supports regular and reentrant wave activity. The defibrillation technique employed involves targeted low-energy perturbations that can be used to “teleport” and/or annihilate singularities present in the excitable BZ medium. 

   more » « less
4. 4D physiologically adaptable cardiac patch: A 4-month in vivo study for the treatment of myocardial infarction

   https://doi.org/10.1126/sciadv.abb5067  

   Cui, Haitao; Liu, Chengyu; Esworthy, Timothy; Huang, Yimin; Yu, Zu-xi; Zhou, Xuan; San, Hong; Lee, Se-jun; Hann, Sung Yun; Boehm, Manfred; et al (June 2020, Science Advances)

   There has been considerable progress in engineering cardiac scaffolds for the treatment of myocardial infarction (MI). However, it is still challenging to replicate the structural specificity and variability of cardiac tissues using traditional bioengineering approaches. In this study, a four-dimensional (4D) cardiac patch with physiological adaptability has been printed by beam-scanning stereolithography. By combining a unique 4D self-morphing capacity with expandable microstructure, the specific design has been shown to improve both the biomechanical properties of the patches themselves and the dynamic integration of the patch with the beating heart. Our results demonstrate improved vascularization and cardiomyocyte maturation in vitro under physiologically relevant mechanical stimulation, as well as increased cell engraftment and vascular supply in a murine chronic MI model. This work not only potentially provides an effective treatment method for MI but also contributes a cutting-edge methodology to enhance the structural design of complex tissues for organ regeneration. 

   more » « less
5. Robot See Robot Do: Imitating Articulated Object Manipulation with Monocular 4D Reconstruction

   Kerr, Justin; Kim, Chung_Min; Wu, Mingxuan; Yi, Brent; Wang, Qianqian; Goldberg, Ken; Kanazawa, Angjoo (November 2024, Conference on Robot Learning)

   Humans can learn to manipulate new objects by simply watching others; providing robots with the ability to learn from such demonstrations would enable a natural interface specifying new behaviors. This work develops Robot See Robot Do (RSRD), a method for imitating articulated object manipulation from a single monocular RGB human demonstration given a single static multi-view object scan. We first propose 4D Differentiable Part Models (4D-DPM), a method for recovering 3D part motion from a monocular video with differentiable rendering. This analysis-by-synthesis approach uses part-centric feature fields in an iterative optimization which enables the use of geometric regularizers to recover 3D motions from only a single video. Given this 4D reconstruction, the robot replicates object trajectories by planning bimanual arm motions that induce the demonstrated object part motion. By representing demonstrations as part-centric trajectories, RSRD focuses on replicating the demonstration's intended behavior while considering the robot's own morphological limits, rather than attempting to reproduce the hand's motion. We evaluate 4D-DPM's 3D tracking accuracy on ground truth annotated 3D part trajectories and RSRD's physical execution performance on 9 objects across 10 trials each on a bimanual YuMi robot. Each phase of RSRD achieves an average of 87% success rate, for a total end-to-end success rate of 60% across 90 trials. Notably, this is accomplished using only feature fields distilled from large pretrained vision models -- without any task-specific training, fine-tuning, dataset collection, or annotation. 

   more » « less