Mr. Junyuan Li Profile

Junyuan Li

at Johns Hopkins Univ

SPIE Involvement:

Author

Publications (8)

Proceedings Article | 1 April 2024 Poster + Paper

Jacobian analysis in deep learning CT denoising networks

Xiaoxuan Zhang, Patrick Li, Junyuan Li, Grace Jianan Gang

Proceedings Volume 12925, 1292525 (2024) https://doi.org/10.1117/12.3006220

KEYWORDS: Denoising, Deep learning, Neural networks, Computed tomography, Lung, Image quality

Read Abstract +

Proceedings Article | 18 October 2022 Open Access

Practical workflow for arbitrary non-circular orbits for CT with clinical robotic C-arms

Yiqun Ma, Grace Gang, Tess Reynolds, Tina Ehitiati, Junyuan Li, Owen Dillon, Tom Russ, Wenying Wang, Clifford Weiss, Nicholas Theodore, Kelvin Hong, Ricky O’Brien, Jeffrey Siewerdsen, J. Webster Stayman

Proceedings Volume 12304, 123042H (2022) https://doi.org/10.1117/12.2647158

KEYWORDS: Calibration, Robotics, Image quality, Spine, Imaging systems, Reconstruction algorithms, Data acquisition, CT reconstruction, X-ray computed tomography

Read Abstract +

DOWNLOAD PAPER

Proceedings Article | 18 October 2022 Open Access

Local linearity analysis of deep learning CT denoising algorithms

Junyuan Li, Wenying Wang, Matthew Tivnan, Jeremias Sulam, Jerry Prince, Michael McNitt-Gray, J. Webster Stayman, Grace Gang

Proceedings Volume 12304, 123040T (2022) https://doi.org/10.1117/12.2646371

KEYWORDS: Neural networks, Denoising, Computed tomography, Complex systems, Algorithm development, Medical imaging, Visualization, Spherical lenses, Space operations, Signal to noise ratio

Read Abstract +

The rapid development of deep-learning methods in medical imaging has called for an analysis method suitable for non-linear and data-dependent algorithms. In this work, we investigate a local linearity analysis where a complex neural network can be represented as piecewise linear systems. We recognize that a large number of neural networks consists of alternating linear layers and rectified linear unit (ReLU) activations, and are therefore strictly piecewise linear. We investigated the extent of these locally linear regions by gradually adding perturbations to an operating point. For this work, we explored perturbations based on image features of interest, including lesion contrast, background, and additive noise. We then developed strategies to extend these strictly locally linear regions to include neighboring linear regions with similar gradients. Using these approximately linear regions, we applied singular value decomposition (SVD) analysis to each local linear system to investigate and explain the overall nonlinear and data-dependent behaviors of neural networks. The analysis was applied to an example CT denoising algorithm trained on thorax CT scans. We observed that the strictly local linear regions are highly sensitive to small signal perturbations. Over a range of lesion contrast from 0.007 to 0.04 mm⁻¹, there is a total of 33992 linear regions. The Jacobians are also shift-variant. However, the Jacobians of neighboring linear regions are very similar. By combining linear regions with similar Jacobians, we narrowed down the number of approximately linear regions to four over lesion contrast from 0.001 to 0.08 mm⁻¹. The SVD analysis to different linear regions revealed denoising behavior that is highly dependent on the background intensity. Analysis further identified greater amount of noise reduction in uniform regions compared to lesion edges. In summary, the local linearity analysis framework we proposed has the potential for us to better characterize and interpret the non-linear and data-dependent behaviors of neural networks.

DOWNLOAD PAPER

Proceedings Article | 4 April 2022 Presentation + Paper

Performance assessment framework for neural network denoising

Junyuan Li, Wenying Wang, Matthew Tivnan, Joseph Stayman, Grace Gang

Proceedings Volume 12031, 1203114 (2022) https://doi.org/10.1117/12.2612732

KEYWORDS: Lung, Denoising, Neural networks, Image quality, Computed tomography

Read Abstract +

The proliferation of deep learning image processing calls for a quantitative image quality assessment framework that is suitable for nonlinear, data-dependent algorithms. In this work, we propose a method to systematically evaluate the system and noise responses such that the nonlinear transfer properties can be mapped out. The method involves sampling of lesion perturbations as a function of size, contrast, as well as clinically relevant features such as shape and texture that may be important for diagnosis. We embed the perturbations in backgrounds of varying attenuation levels, noise magnitude and correlation that are associated with different patient anatomies and imaging protocols. The range of system and noise response are further used to evaluate performance for clinical tasks such as signal detection and classification. We performed the assessment for an example CNN-denoising algorithm for low does lung CT screening. The system response of the CNN-denoising algorithm exhibits highly nonlinear behavior where both contrast and higher order lesion features such as spiculated boundaries are not reliably represented for lesions perturbations with small size and low contrast. The noise properties are potentially highly nonstationary, and should be assumed to be the same between the signal-present and signal-absent images. Furthermore, we observer a high degree dependency of both system and noise response on the background attenuation levels. Inputs around zeros are effectively imposed a non-negativity constraint; transfer properties for higher background levels are highly variable. For a detection task, CNN-denoised images improved detectability index by 16-18% compared to low dose CT inputs. For classification task between spiculated and smooth lesions, CNN-denoised images result in a much larger improvement up to 50%. The performance assessment framework propose in this work can systematically map out the nonlinear transfer functions for deep learning algorithms and can potentially enable robust deployment of such algorithms in medical imaging applications.

Proceedings Article | 4 April 2022 Presentation + Paper

Data-dependent nonlinearity analysis in CT denoising CNNs

Wenying Wang, Matthew Tivnan, Junyuan Li, Joseph Stayman, Grace Gang

Proceedings Volume 12031, 1203117 (2022) https://doi.org/10.1117/12.2612569

KEYWORDS: Neural networks, Denoising, Computed tomography

Read Abstract +

Proceedings Article | 4 April 2022 Presentation + Paper

Assessment of boundary discrimination performance in a printed phantom

Craig Abbey, Junyuan Li, Grace Gang, J. Webster Stayman

Proceedings Volume 12035, 120350N (2022) https://doi.org/10.1117/12.2612622

KEYWORDS: Image quality, Computed tomography, Performance modeling, Imaging systems

Read Abstract +

Proceedings Article | 15 February 2021 Poster + Presentation

Mitigating unknown biases in CT data using machine learning

Junyuan Li, Grace Gang, Webster Stayman

Proceedings Volume 11595, 1159541 (2021) https://doi.org/10.1117/12.2582274

KEYWORDS: Machine learning, Systems modeling, Data processing, Sensors, Model-based design, Image processing, Data modeling, Computed tomography, CT reconstruction, Physics

Read Abstract +

Proceedings Article | 16 March 2020 Presentation + Paper

Performance assessment of texture reproduction in high-resolution CT

Hui Shi, Grace Gang, Junyuan Li, Eleni Liapi, Craig Abbey, J. Webster Stayman

Proceedings Volume 11316, 113160R (2020) https://doi.org/10.1117/12.2550579

KEYWORDS: Computed tomography, Spatial resolution, Diagnostics, Scanners, Image resolution, Image quality

Read Abstract +

Showing 5 of 8 publications

View contact details

UPDATE YOUR PROFILE

Is this your profile? Update it now.

Sign into your SPIE.org account

Don’t have a profile and want one?

Create an account on SPIE.org

Keywords/Phrases

Search In:

Publication Years