Dr. Joshua Victor Stough Profile

Dr. Joshua Victor Stough

at Bucknell Univ

SPIE Involvement:

Conference Program Committee | Author

Area of Expertise:

medical image analysis , deep learning , computer vision , image processing , segmentation

Websites:

Company Website

Dr. Stough's research leverages machine learning and high-performance computing to help automate medical image analysis. Examples include planning prostate radiation treatment in computed tomography (CT), and analyzing sub-cortical brain structures in magnetic resonance and diffusion tensor imaging (MR-DTI). Recently his work has focused on quantifying heart structure and function in cardiac MR and echocardiography, in partnership with the regional Geisinger Health hospital system. The group is focused on the development of machine learning and image processing techniques to advance precision medicine in the study of cardiovascular disease.

Proceedings Article | 3 April 2023 Presentation + Paper

Co-Unet-GAN: a co-learning domain adaptation model on echocardiography segmentation

Junyang Cai, Christopher Haggerty, Joshua Stough

Proceedings Volume 12464, 124641B (2023) https://doi.org/10.1117/12.2652621

KEYWORDS: Image segmentation, Education and training, Data modeling, Performance modeling, Gallium nitride, Echocardiography, Medical imaging, Image restoration, Machine learning, Process modeling

Read Abstract +

Proceedings Article | 4 April 2022 Poster + Paper

Perceptually improved T1-T2 MRI translations using conditional generative adversarial networks

Anurag Vaidya, Joshua Stough, Aalpen Patel

Proceedings Volume 12032, 120321V (2022) https://doi.org/10.1117/12.2608428

KEYWORDS: Data modeling, Magnetic resonance imaging, Brain, Convolution, Neuroimaging, Image quality, Tumors, Visualization, Performance modeling

Read Abstract +

Proceedings Article | 4 April 2022 Presentation + Paper

Fully automated multi-heartbeat echocardiography video segmentation and motion tracking

Yida Chen, Xiaoyan Zhang, Christopher Haggerty, Joshua Stough

Proceedings Volume 12032, 120320R (2022) https://doi.org/10.1117/12.2607871

KEYWORDS: Video, Echocardiography, Image segmentation, Heart, Motion analysis, Convolution, Feature extraction

Read Abstract +

Proceedings Article | 15 February 2021 Presentation + Paper

Assessing the generalizability of temporally coherent echocardiography video segmentation

Yida Chen, Xiaoyan Zhang, Christopher Haggerty, Joshua Stough

Proceedings Volume 11596, 115961O (2021) https://doi.org/10.1117/12.2580874

KEYWORDS: Video, Echocardiography, Heart, Temporal coherence, Reliability, Data modeling

Read Abstract +

Proceedings Article | 10 March 2020 Presentation + Paper

Left ventricular and atrial segmentation of 2D echocardiography with convolutional neural networks

Joshua Stough, Sushravya Raghunath, Xiaoyan Zhang, John Pfeifer, Brandon Fornwalt, Christopher Haggerty

Proceedings Volume 11313, 113130A (2020) https://doi.org/10.1117/12.2547375

KEYWORDS: Echocardiography, Image segmentation, Convolutional neural networks, Computer programming, Transducers, Heart, Image quality, Binary data, Computer science, Machine learning

Read Abstract +

Proceedings Article | 12 March 2018 Paper

Ventricular segmentation and quantitative assessment in cardiac MR using convolutional neural networks

Joshua Stough, Joseph DiPalma, Zilin Ma, Brandon Fornwalt, Christopher Haggerty

Proceedings Volume 10578, 1057826 (2018) https://doi.org/10.1117/12.2291534

KEYWORDS: Cardiovascular magnetic resonance imaging, Image segmentation, Convolutional neural networks, Heart, Neural networks, Data modeling, Magnetic resonance imaging, Image processing, Blood, Statistical modeling

Read Abstract +

Proceedings Article | 21 March 2016 Paper

Thalamus parcellation using multi-modal feature classification and thalamic nuclei priors

Jeffrey Glaister, Aaron Carass, Joshua Stough, Peter Calabresi, Jerry Prince

Proceedings Volume 9784, 97843J (2016) https://doi.org/10.1117/12.2216987

KEYWORDS: Thalamus, Image segmentation, Image processing algorithms and systems, Diffusion magnetic resonance imaging, Magnetic resonance imaging, Machine learning, Magnetism, Algorithm development, Diffusion, Diffusion tensor imaging, Brain, Neuroimaging, Feature extraction

Read Abstract +

Proceedings Article | 26 March 2008 Paper

Conditional-mean initialization using neighboring objects in deformable model segmentation

Ja-Yeon Jeong, Joshua Stough, J. Steve Marron, Stephen Pizer

Proceedings Volume 6914, 69144R (2008) https://doi.org/10.1117/12.770712

KEYWORDS: Image segmentation, Chemical species, Prostate, Bladder, Rectum, Statistical analysis, Shape analysis, Principal component analysis, Statistical modeling, Data modeling

Read Abstract +

Most model-based segmentation methods find a target object in a new image by constructing an objective function and optimizing it using a standard minimization algorithm. In general, the objective function has two penalty terms: 1) for deforming a template model and 2) for mismatch between the trained image intensities relative to the template model and the observed image intensities relative to the deformed template model in the target image. While it is difficult to establish an objective function with a global minimum at the desired segmentation result, even such an objective function is typically non-convex due to the complexity of the intensity patterns and the many structures surrounding the target object. Thus, it is critical that the optimization starts at a point close to the global minimum of the objective function in deformable model-based segmentation framework. For a segmentation method in maximum a posteriori framework a good objective function can be obtained by learning the probability distributions of the population shape deformations and their associated image intensities because each penalty term can be simplified to a squared function of some distance metric defined in the shape space. The mean shape and intensities of the learned probability distributions also provide a good initialization for segmentation. However, a major concern in estimating the shape prior is the stability of the estimated shape distributions from given training samples because the feature space of a shape model is usually very high dimensional while the number of training samples is limited. A lot of effort in that regard have been made to attain a stable estimation of shape probability distribution. In this paper, we describe our approach to stably estimate a shape probability distribution when good segmentations of objects adjacent to the target object are available. Our approach is to use a conditional shape probability distribution (CSPD) to take into account in the shape distribution the relation of the target object to neighboring objects. In particular, we propose a new method based on principal component regression (PCR) in reflecting in the conditional term of the CSPD the effect of neighboring objects on the target object. The resulting approach is able to give a better and robust initialization with training samples of a few cases. To demonstrate the potential of our approach, we apply it first to training of a simulated data of known deformations and second to male pelvic organs, using the CSPD in m-rep segmentations of the prostate in CT images. Our results show a clear improvement in initializing the prostate by its conditional mean given the bladder and the rectum as neighboring objects, as measured both by volume overlap and average surface distance.

Showing 5 of 8 publications

Conference Committee Involvement (3)

Image Processing

17 February 2025 | San Diego, California, United States

Image Processing

19 February 2024 | San Diego, California, United States

Image Processing

20 February 2023 | San Diego, California, United States

Keywords/Phrases

Search In:

Publication Years