arXiv subjects

Lian Zhang

Publications and source records attributed to Lian Zhang.

48 records · Page 3Linked to original sources

arXiv2023

Segment Anything Model (SAM) for Radiation Oncology

In this study, we evaluate the performance of the Segment Anything Model (SAM) in clinical radiotherapy. Our results indicate that SAM's 'segment anything' mode can achieve clinically acceptable segmentation results in most organs-at-risk (OARs) with Dice scores higher than 0.7. SAM's 'box prompt' mode further improves the Dice scores by 0.1 to 0.5. Considering the size of the organ and the clarity of its boundary, SAM displays better performance for large organs with clear boundaries but performs worse for smaller organs with unclear boundaries. Given that SAM, a model pre-trained purely on natural images, can handle the delineation of OARs from medical images with clinically acceptable accuracy, these results highlight SAM's robust generalization capabilities with consistent accuracy in automatic segmentation for radiotherapy. In other words, SAM can achieve delineation of different OARs at different sites using a generic automatic segmentation model. SAM's generalization capabilities across different disease sites suggest that it is technically feasible to develop a generic model for automatic segmentation in radiotherapy.

eess.IV↗

arXiv2023

Beam mask and sliding window-facilitated deep learning-based accurate and efficient dose prediction for pencil beam scanning proton therapy

Purpose: To develop a DL-based PBSPT dose prediction workflow with high accuracy and balanced complexity to support on-line adaptive proton therapy clinical decision and subsequent replanning. Methods: PBSPT plans of 103 prostate cancer patients and 83 lung cancer patients previously treated at our institution were included in the study, each with CTs, structure sets, and plan doses calculated by the in-house developed Monte-Carlo dose engine. For the ablation study, we designed three experiments corresponding to the following three methods: 1) Experiment 1, the conventional region of interest (ROI) method. 2) Experiment 2, the beam mask (generated by raytracing of proton beams) method to improve proton dose prediction. 3) Experiment 3, the sliding window method for the model to focus on local details to further improve proton dose prediction. A fully connected 3D-Unet was adopted as the backbone. Dose volume histogram (DVH) indices, 3D Gamma passing rates, and dice coefficients for the structures enclosed by the iso-dose lines between the predicted and the ground truth doses were used as the evaluation metrics. The calculation time for each proton dose prediction was recorded to evaluate the method's efficiency. Results: Compared to the conventional ROI method, the beam mask method improved the agreement of DVH indices for both targets and OARs and the sliding window method further improved the agreement of the DVH indices. For the 3D Gamma passing rates in the target, OARs, and BODY (outside target and OARs), the beam mask method can improve the passing rates in these regions and the sliding window method further improved them. A similar trend was also observed for the dice coefficients. In fact, this trend was especially remarkable for relatively low prescription isodose lines. The dose predictions for all the testing cases were completed within 0.25s.

physics.med-ph↗

arXiv2023

Evaluating Large Language Models on a Highly-specialized Topic, Radiation Oncology Physics

We present the first study to investigate Large Language Models (LLMs) in answering radiation oncology physics questions. Because popular exams like AP Physics, LSAT, and GRE have large test-taker populations and ample test preparation resources in circulation, they may not allow for accurately assessing the true potential of LLMs. This paper proposes evaluating LLMs on a highly-specialized topic, radiation oncology physics, which may be more pertinent to scientific and medical communities in addition to being a valuable benchmark of LLMs. We developed an exam consisting of 100 radiation oncology physics questions based on our expertise at Mayo Clinic. Four LLMs, ChatGPT (GPT-3.5), ChatGPT (GPT-4), Bard (LaMDA), and BLOOMZ, were evaluated against medical physicists and non-experts. ChatGPT (GPT-4) outperformed all other LLMs as well as medical physicists, on average. The performance of ChatGPT (GPT-4) was further improved when prompted to explain first, then answer. ChatGPT (GPT-3.5 and GPT-4) showed a high level of consistency in its answer choices across a number of trials, whether correct or incorrect, a characteristic that was not observed in the human test groups. In evaluating ChatGPTs (GPT-4) deductive reasoning ability using a novel approach (substituting the correct answer with "None of the above choices is the correct answer."), ChatGPT (GPT-4) demonstrated surprising accuracy, suggesting the potential presence of an emergent ability. Finally, although ChatGPT (GPT-4) performed well overall, its intrinsic properties did not allow for further improvement when scoring based on a majority vote across trials. In contrast, a team of medical physicists were able to greatly outperform ChatGPT (GPT-4) using a majority vote. This study suggests a great potential for LLMs to work alongside radiation oncology experts as highly knowledgeable assistants.

physics.med-ph↗

arXiv2023

FV-MgNet: Fully Connected V-cycle MgNet for Interpretable Time Series Forecasting

By investigating iterative methods for a constrained linear model, we propose a new class of fully connected V-cycle MgNet for long-term time series forecasting, which is one of the most difficult tasks in forecasting. MgNet is a CNN model that was proposed for image classification based on the multigrid (MG) methods for solving discretized partial differential equations (PDEs). We replace the convolutional operations with fully connected operations in the existing MgNet and then apply them to forecasting problems. Motivated by the V-cycle structure in MG, we further propose the FV-MgNet, a V-cycle version of the fully connected MgNet, to extract features hierarchically. By evaluating the performance of FV-MgNet on popular data sets and comparing it with state-of-the-art models, we show that the FV-MgNet achieves better results with less memory usage and faster inference speed. In addition, we develop ablation experiments to demonstrate that the structure of FV-MgNet is the best choice among the many variants.

cs.AI↗

arXiv2022

An Interpretive Constrained Linear Model for ResNet and MgNet

We propose a constrained linear data-feature-mapping model as an interpretable mathematical model for image classification using a convolutional neural network (CNN). From this viewpoint, we establish detailed connections between the traditional iterative schemes for linear systems and the architectures of the basic blocks of ResNet- and MgNet-type models. Using these connections, we present some modified ResNet models that compared with the original models have fewer parameters and yet can produce more accurate results, thereby demonstrating the validity of this constrained learning data-feature-mapping assumption. Based on this assumption, we further propose a general data-feature iterative scheme to show the rationality of MgNet. We also provide a systematic numerical study on MgNet to show its success and advantages in image classification problems and demonstrate its advantages in comparison with established networks.

cs.CV↗

arXiv2022

Predicting Alzheimer's Disease Using 3DMgNet

Alzheimer's disease (AD) is an irreversible neurode generative disease of the brain.The disease may causes memory loss, difficulty communicating and disorientation. For the diagnosis of Alzheimer's disease, a series of scales are often needed to evaluate the diagnosis clinically, which not only increases the workload of doctors, but also makes the results of diagnosis highly subjective. Therefore, for Alzheimer's disease, imaging means to find early diagnostic markers has become a top priority. In this paper, we propose a novel 3DMgNet architecture which is a unified framework of multigrid and convolutional neural network to diagnose Alzheimer's disease (AD). The model is trained using an open dataset (ADNI dataset) and then test with a smaller dataset of ours. Finally, the model achieved 92.133% accuracy for AD vs NC classification and significantly reduced the model parameters.

eess.IV↗

arXiv2020

Constrained Linear Data-feature Mapping for Image Classification

In this paper, we propose a constrained linear data-feature mapping model as an interpretable mathematical model for image classification using convolutional neural network (CNN) such as the ResNet. From this viewpoint, we establish the detailed connections in a technical level between the traditional iterative schemes for constrained linear system and the architecture for the basic blocks of ResNet. Under these connections, we propose some natural modifications of ResNet type models which will have less parameters but still maintain almost the same accuracy as these corresponding original models. Some numerical experiments are shown to demonstrate the validity of this constrained learning data-feature mapping assumption.

eess.IV↗

arXiv2020

Make $\ell_1$ Regularization Effective in Training Sparse CNN

Compressed Sensing using $\ell_1$ regularization is among the most powerful and popular sparsification technique in many applications, but why has it not been used to obtain sparse deep learning model such as convolutional neural network (CNN)? This paper is aimed to provide an answer to this question and to show how to make it work. We first demonstrate that the commonly used stochastic gradient decent (SGD) and variants training algorithm is not an appropriate match with $\ell_1$ regularization and then replace it with a different training algorithm based on a regularized dual averaging (RDA) method. RDA was originally designed specifically for convex problem, but with new theoretical insight and algorithmic modifications (using proper initialization and adaptivity), we have made it an effective match with $\ell_1$ regularization to achieve a state-of-the-art sparsity for CNN compared to other weight pruning methods without compromising accuracy (achieving 95\% sparsity for ResNet18 on CIFAR-10, for example).

cs.LG↗

arXiv2019

Decoupling PDE Computation with Intrinsic or Inertial Robin Interface Condition

We study decoupled numerical methods for multi-domain, multi-physics applications. By investigating various stages of numerical approximation and decoupling and tracking how the information is transmitted across the interface for a typical multi-modeling model problem, we derive an approximate intrinsic or inertial type Robin condition for its semi-discrete model. This new interface condition is justified both mathematically and physically in contrast to the classical Robin interface condition conventionally introduced for decoupling multi-modeling problems. Based on the intrinsic or inertial Robin condition, an equivalent semi-discrete model is introduced, which provides a general framework for devising effective decoupled numerical methods. Numerical experiments also confirm the effectiveness of this new decoupling approach.

math.NA↗

arXiv2019

Parallaxes for star forming regions in the inner Perseus spiral arm

We report trigonometric parallax and proper motion measurements of 6.7-GHz CH3OH and 22-GHz H2O masers in eight high-mass star-forming regions (HMSFRs) based on VLBA observations as part of the BeSSeL Survey. The distances of these HMSFRs combined with their Galactic coordinates, radial velocities, and proper motions, allow us to assign them to a segment of the Perseus arm with ~< 70 deg. These HMSFRs are clustered in Galactic longitude from ~30 deg to ~50, neighboring a dirth of such sources between longitudes ~50 deg to ~90 deg.

astro-ph.GA↗

arXiv2018

A Multirate Approach for Fluid-Structure Interaction Computation with Decoupled Methods

We investigate a multirate time step approach applied to decoupled methods in fluid and structure interaction(FSI) computation, where two different time steps are used for fluid and structure respectively. For illustration, the multirate technique is tested by the decoupled β-scheme. Numerical experiments show that the proposed approach is stable and retains the same order accuracy as the original single time step schemes, while with much less computational expense.

math.NA↗

arXiv2009

Magnetic cooling at Risoe DTU

Magnetic refrigeration at room temperature is of great interest due to a long-term goal of making refrigeration more energy-efficient, less noisy and free of any environmentally hostile materials. A refrigerator utilizing an active magnetic regenerator (AMR) is based on the magnetocaloric effect, which manifests itself as a temperature change in magnetic materials when subjected to a varying magnetic field. In this work we present the current state of magnetic refrigeration research at Risoe DTU with emphasis on the numerical modeling of an existing AMR test machine. A 2D numerical heat-transfer and fluid-flow model that represents the experimental setup is presented. Experimental data of both no-heat load and heat load situations are compared to the model. Moreover, results from the numerical modeling of the permanent magnet design used in the system are presented.

cond-mat.mtrl-sci↗