BZPEERReading, trust and collaboration for research
Menu

Discover

GraphFollow connections around a paper, topic or saved view.

Workspaces

HomePick up where your research flow left off.InboxHandle requests, replies and notifications from one place.CollectionsCurate reading lists, evidence sets and shareable dossiers.ResearchSee your private provenance graph, trail and imports.CollaborationMove from discovery into focused discussion rooms.PublishDraft, preview and publish full-text research articles.

Network

ExpertsFind specialists worth following or asking for help.CommunitiesJoin research circles organized around shared work.PartnersSee external organizations active around the same topics.

Opportunities

ListingsBrowse roles, calls and collaborations with clearer fit signals.

Account

Log inReturn to your reading and collaboration workspace.Create accountStart saving work, following people and joining teams.
Log inCreate account

Researcher profile

Lei Xing

Lei Xing contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate
Computer Visioneess.IVphysics.med-phMachine Learningmath.OCastro-ph.EPComputational Engineering, Finance, and Sciencemath.STphysics.app-phQuantitative MethodsStatistics Theory

Trust snapshot

Quick read

Trust 21 - EmergingVerification L1Unclaimed author
23works
0followers
11topics
4close collaborators

Actions

Decide how to stay connected

Follow researcher0

Identity and collaboration

How to connect with this researcher

Claiming links this public author record to a researcher profile and unlocks direct collaboration workflows.

Log in to claim

Direct collaboration

Open a focused conversation when the fit is right

Claim this author entity first to unlock direct invitations.

Research graph

See the researcher in context

Open full explorer

Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.

Building this graph slice

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Published work

23 published item(s)

preprint2026arXiv

Computational Imaging Priors for Wireless Capsule Endoscopy: Monte Carlo-Guided Hemoglobin Mapping for Rare-Anomaly Detection

Background. RGB-trained capsule-endoscopy classifiers underperform on small-vessel vascular findings by conflating hemoglobin contrast with bile and illumination falloff. Thus, here we test whether a Monte Carlo-inspired analytic model can compute hemoglobin from RGB signal built upon extracted classifier. Methods. On Kvasir-Capsule (47,238 frames, video-level 70/15/15 split, 11 evaluable classes) we evaluate two software-only configurations against RGB-only EfficientNet-B0 across 6 seeds: (i) a prior P_blood = sigma(alpha * (H_norm - 0.5)) * Phi(r) fused as 2 zero-init auxiliary channels; (ii) a distillation head training a 3-channel RGB backbone to predict P_blood. Significance: paired DeLong, McNemar, bootstrap CIs with Bonferroni correction. Results. Across 6 seeds (n=6,423), the analytic prior provides a small but direction-consistent macro-AUC improvement: RGB-only 0.760 +/- 0.027, input-fusion 0.783 +/- 0.024 (paired Delta = +0.023, sign-positive on 5/6 seeds), distillation 0.773 +/- 0.028. The largest robust per-class lift is on Lymphangiectasia, where AUC rises from RGB 0.238 +/- 0.057 to input-fusion 0.337 +/- 0.019, sign-consistent across all 6 seeds. On rare focal-vascular classes (Angiectasia, Blood - fresh) the prior's per-seed effects are bimodal: seed=42 reaches Angiectasia AUC 0.528 -> 0.916, but the cross-seed mean is 0.646 -> 0.608 with sigma_PI = 0.23 - reported as a high-variance per-seed exemplar. Conclusion. A Monte Carlo-inspired analytic prior provides a small, direction-consistent macro-AUC improvement on Kvasir-Capsule across 6 seeds with the largest robust per-class lift on Lymphangiectasia; the distillation variant runs on plain 3-channel RGB and yields a free interpretability heatmap.

0 citations0 reviewsNo code linkedOpen access
preprint2023arXiv

Label-Efficient Self-Supervised Federated Learning for Tackling Data Heterogeneity in Medical Imaging

The collection and curation of large-scale medical datasets from multiple institutions is essential for training accurate deep learning models, but privacy concerns often hinder data sharing. Federated learning (FL) is a promising solution that enables privacy-preserving collaborative learning among different institutions, but it generally suffers from performance deterioration due to heterogeneous data distributions and a lack of quality labeled data. In this paper, we present a robust and label-efficient self-supervised FL framework for medical image analysis. Our method introduces a novel Transformer-based self-supervised pre-training paradigm that pre-trains models directly on decentralized target task datasets using masked image modeling, to facilitate more robust representation learning on heterogeneous data and effective knowledge transfer to downstream models. Extensive empirical results on simulated and real-world medical imaging non-IID federated datasets show that masked image modeling with Transformers significantly improves the robustness of models against various degrees of data heterogeneity. Notably, under severe data heterogeneity, our method, without relying on any additional pre-training data, achieves an improvement of 5.06%, 1.53% and 4.58% in test accuracy on retinal, dermatology and chest X-ray classification compared to the supervised baseline with ImageNet pre-training. In addition, we show that our federated self-supervised pre-training methods yield models that generalize better to out-of-distribution data and perform more effectively when fine-tuning with limited labeled data, compared to existing FL algorithms. The code is available at https://github.com/rui-yan/SSL-FL.

0 citations0 reviewsNo code linkedOpen access
preprint2023arXiv

NeRP: Implicit Neural Representation Learning with Prior Embedding for Sparsely Sampled Image Reconstruction

Image reconstruction is an inverse problem that solves for a computational image based on sampled sensor measurement. Sparsely sampled image reconstruction poses addition challenges due to limited measurements. In this work, we propose an implicit Neural Representation learning methodology with Prior embedding (NeRP) to reconstruct a computational image from sparsely sampled measurements. The method differs fundamentally from previous deep learning-based image reconstruction approaches in that NeRP exploits the internal information in an image prior, and the physics of the sparsely sampled measurements to produce a representation of the unknown subject. No large-scale data is required to train the NeRP except for a prior image and sparsely sampled measurements. In addition, we demonstrate that NeRP is a general methodology that generalizes to different imaging modalities such as CT and MRI. We also show that NeRP can robustly capture the subtle yet significant image changes required for assessing tumor progression.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

CateNorm: Categorical Normalization for Robust Medical Image Segmentation

Batch normalization (BN) uniformly shifts and scales the activations based on the statistics of a batch of images. However, the intensity distribution of the background pixels often dominates the BN statistics because the background accounts for a large proportion of the entire image. This paper focuses on enhancing BN with the intensity distribution of foreground pixels, the one that really matters for image segmentation. We propose a new normalization strategy, named categorical normalization (CateNorm), to normalize the activations according to categorical statistics. The categorical statistics are obtained by dynamically modulating specific regions in an image that belong to the foreground. CateNorm demonstrates both precise and robust segmentation results across five public datasets obtained from different domains, covering complex and variable data distributions. It is attributable to the ability of CateNorm to capture domain-invariant information from multiple domains (institutions) of medical data. Code is available at https://github.com/lambert-x/CateNorm.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Image Classification using Graph Neural Network and Multiscale Wavelet Superpixels

Prior studies using graph neural networks (GNNs) for image classification have focused on graphs generated from a regular grid of pixels or similar-sized superpixels. In the latter, a single target number of superpixels is defined for an entire dataset irrespective of differences across images and their intrinsic multiscale structure. On the contrary, this study investigates image classification using graphs generated from an image-specific number of multiscale superpixels. We propose WaveMesh, a new wavelet-based superpixeling algorithm, where the number and sizes of superpixels in an image are systematically computed based on its content. WaveMesh superpixel graphs are structurally different from similar-sized superpixel graphs. We use SplineCNN, a state-of-the-art network for image graph classification, to compare WaveMesh and similar-sized superpixels. Using SplineCNN, we perform extensive experiments on three benchmark datasets under three local-pooling settings: 1) no pooling, 2) GraclusPool, and 3) WavePool, a novel spatially heterogeneous pooling scheme tailored to WaveMesh superpixels. Our experiments demonstrate that SplineCNN learns from multiscale WaveMesh superpixels on-par with similar-sized superpixels. In all WaveMesh experiments, GraclusPool performs poorer than no pooling / WavePool, indicating that poor choice of pooling can result in inferior performance while learning from multiscale superpixels.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Operator Splitting for Adaptive Radiation Therapy with Nonlinear Health Dynamics

We present an optimization-based approach to radiation treatment planning over time. Our approach formulates treatment planning as an optimal control problem with nonlinear patient health dynamics derived from the standard linear-quadratic cell survival model. As the formulation is nonconvex, we propose a method for obtaining an approximate solution by solving a sequence of convex optimization problems. This method is fast, efficient, and robust to model error, adapting readily to changes in the patient's health between treatment sessions. Moreover, we show that it can be combined with the operator splitting method ADMM to produce an algorithm that is highly scalable and can handle large clinical cases. We introduce an open-source Python implementation of our algorithm, AdaRad, and demonstrate its performance on several examples.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Patient-specific mean teacher UNet for enhancing PET image and low-dose PET reconstruction on RefleXion X1 biology-guided radiotherapy system

The RefleXion X1 is the first biology-guided radiotherapy (BgRT) system. Its dual 90-degree PET detector collects fewer pair production events compared to a full-ring diagnostic PET system. In the proposed BgRT workflow, a short scan is acquired before treatment delivery to ensure image quality and consistency. The shorter scan time, a quarter of the simulation scan time, also leads to fewer coincidence events and hence reduced image quality. In this study, we proposed a patient-specific mean teacher UNet (MT-UNet) to enhance PET image quality and low-dose PET reconstruction on RefleXion X1. PET/CT scans of nine cancer patients were acquired using RefleXion X1. Every patient had one simulation scan. Five patients had additional scans acquired during the first and the final treatment fractions. Treatment scans were acquired using the same imaging protocol as the simulation scan. For each scan, we reconstructed a full-dose image and evenly split coincidence events into four sessions to reconstruct four quarter-dose PET images. For each patient, our proposed MT-UNet was trained using quarter-dose and full-dose images of the simulation scan. For the image quality enhancement task, we applied nine trained MT-UNets to full-dose simulation PET images of the nine patients to generate enhanced images, respectively. The enhanced images were compared with the original full-dose images using CNR and SNR. For the low-dose image reconstruction task, we applied five trained MT-UNets to ten quarter-dose treatment images of five patients to predict full-dose images, respectively. The predicted and ground truth full-dose images were compared using SSIM and PSNR. We also trained and evaluated patient-specific UNets for model comparison. Our proposed patient-specific MT-UNet achieved better performance in improving the quality of RefleXion low-dose and full-dose images compared to the patient-specific UNet.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Solving Inverse Problems in Medical Imaging with Score-Based Generative Models

Reconstructing medical images from partial measurements is an important inverse problem in Computed Tomography (CT) and Magnetic Resonance Imaging (MRI). Existing solutions based on machine learning typically train a model to directly map measurements to medical images, leveraging a training dataset of paired images and measurements. These measurements are typically synthesized from images using a fixed physical model of the measurement process, which hinders the generalization capability of models to unknown measurement processes. To address this issue, we propose a fully unsupervised technique for inverse problem solving, leveraging the recently introduced score-based generative models. Specifically, we first train a score-based generative model on medical images to capture their prior distribution. Given measurements and a physical model of the measurement process at test time, we introduce a sampling method to reconstruct an image consistent with both the prior and the observed measurements. Our method does not assume a fixed measurement process during training, and can thus be flexibly adapted to different measurement processes at test time. Empirically, we observe comparable or better performance to supervised learning techniques in several medical imaging tasks in CT and MRI, while demonstrating significantly better generalization to unknown measurement processes.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

A Slot Antenna Array with Reconfigurable RCS Using Liquid Absorber

This paper presents a slot antenna array with a reconfigurable radar cross section (RCS). The antenna system is formed by combining a liquid absorber with a 2*2 slot antenna array. The liquid absorber consists of a polymethyl methacrylate (PMMA) container, a 45% ethanol layer, and a metal ground,which is attached to the surface of the slot antenna array. The incident wave can be absorbed by the absorber rather than reflected in other directions when the PMMA container is filled with ethanol, which reduces the monostatic and bistatic RCS. Thus the RCS of the antenna can be changed by injecting and extracting ethanol while the antenna's radiation performance in terms of bandwidth, radiation patterns and gain is well sustained. In a complex communication system, this can be used to switch between detection and stealth mode. The mechanism of the absorber is investigated. The simulated results show that the antenna with this absorber has monostatic and bistatic RCS reduction bands from 2.0 GHz to 18.0 GHz, a maximum RCS reduction of 35 dB with an average RCS reduction of 13.28 dB. The antenna's operating band is 100 MHz. Without ethanol, the antenna has a realized gain of 12.1 dBi, and it drops by 2 dB when the lossy ethanol is injected. The measured results agree well with the simulated ones.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Atmosphere escape inferred from modelling the H$α$ transmission spectrum of WASP-121b

The escaping atmospheres of hydrogen driven by stellar X-ray and extreme Ultraviolet (XUV) have been detected around some exoplanets by the excess absorption of Ly$α$ in far ultraviolet band. In the optical band the excess absorption of H$α$ is also found by the ground-based instruments. However, it is not certain so far if the escape of the atmosphere driven by XUV can result in such absorption. Here we present the XUV driven hydrodynamic simulation coupled with the calculation of detailed level population and the process of radiative transfer for WASP-121b. Our fiducial model predicts a mass loss rate of $\sim$1.28$\times$10$^{12}$g/s for WASP-121b. Due to the high temperature and Ly$α$ intensity predicted by the fiducial model, many hydrogen atoms are populated into the first excited state. As a consequence, the transmission spectrum of H$α$ simulated by our model is broadly consistent with the observation. Comparing with the absorption of H$α$ at different observation times, the stellar XUV emission varies in the range of 0.5-1.5 times fiducial value, which may reflect the variation of the stellar activity. Finally, we find that the supersonic regions of the planetary wind contribute a prominent portion to the absorption of H$α$ by comparing the equivalent width of H$α$, which hints that a transonic outflow of the upper atmosphere driven by XUV irradiation of the host star can be detected by the ground-based telescope and the H$α$ can be a good indicator of escaping atmosphere.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Fully Automated Noncoplanar Radiation Therapy Treatment Planning

Noncoplanar radiation therapy treatment planning has the potential to improve dosimetric quality as compared to traditional coplanar techniques. Likewise, automated treatment planning algorithms can reduce a planner's active treatment planning time and remove inter-planner variability. To address the limitations of traditional treatment planning, we have been developing a suite of algorithms called station parameter optimized radiation therapy (SPORT). Within the SPORT suite of algorithms, we propose a method called NC-POPS to produce noncoplanar (NC) plans using the fully automated Pareto Optimal Projection Search (POPS) algorithm. Our NC-POPS algorithm extends the original POPS algorithm to the noncoplanar setting with potential applications to both IMRT and VMAT. The proposed algorithm consists of two main parts: 1) noncoplanar beam angle optimization (BAO) and 2) fully automated inverse planning using the POPS algorithm. We evaluate the performance of NC-POPS by comparing between various noncoplanar and coplanar configurations. To evaluate plan quality, we compute the homogeneity index (HI), conformity index (CI), and dose-volume histogram (DVH) statistics for various organs-at-risk (OARs). As compared to the evaluated coplanar baseline methods, the proposed NC-POPS method achieves significantly better OAR sparing, comparable or better dose conformity, and similar dose homogeneity. Our proposed NC-POPS algorithm provides a modular approach for fully automated treatment planning of noncoplanar IMRT cases with the potential to substantially improve treatment planning workflow and plan quality.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Pareto Optimal Projection Search (POPS): Automated Radiation Therapy Treatment Planning by Direct Search of the Pareto Surface

Objective: Radiation therapy treatment planning is a time-consuming, iterative process with potentially high inter-planner variability. Fully automated treatment planning processes could reduce a planner's active treatment planning time and remove inter-planner variability, with the potential to tremendously improve patient turnover and quality of care. In developing fully automated algorithms for treatment planning, we have two main objectives: to produce plans that are 1) Pareto optimal and 2) clinically acceptable. Here, we propose the Pareto optimal projection search (POPS) algorithm, which provides a general framework for directly searching the Pareto front. Methods: Our POPS algorithm is a novel automated planning method that combines two main search processes: 1) gradient-free search in the decision variable space and 2) projection of decision variables to the Pareto front using the bisection method. We demonstrate the performance of POPS by comparing with clinical treatment plans. As one possible quantitative measure of treatment plan quality, we construct a clinical acceptability scoring function (SF) modified from the previously developed general evaluation metric (GEM). Results: On a dataset of 21 prostate cases collected as part of clinical workflow, our proposed POPS algorithm produces Pareto optimal plans that are clinically acceptable in regards to dose conformity, dose homogeneity, and sparing of organs-at-risk. Conclusion: Our proposed POPS algorithm provides a general framework for fully automated treatment planning that achieves clinically acceptable dosimetric quality without requiring active planning from human planners. Significance: Our fully automated POPS algorithm addresses many key limitations of other automated planning approaches, and we anticipate that it will substantially improve treatment planning workflow.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

A deep learning approach for virtual monochromatic spectral CT imaging with a standard single energy CT scanner

Purpose/Objectives: To develop and assess a strategy of using deep learning (DL) to generate virtual monochromatic CT (VMCT) images from a single-energy CT (SECT) scan. Materials/Methods: The proposed data-driven VMCT imaging consists of two steps: (i) using a supervised DL model trained with a large number of 100 kV and 140 kV dual-energy CT (DECT) image pairs to produce the corresponding high-energy CT image from a low-energy image; and (ii) reconstructing VMCT images with energy ranging from 40 to 150 keV. To evaluate the performance of the method, we retrospectively studied 6,767 abdominal DECT images. The VMCT images reconstructed using both DL-derived DECT (DL-DECT) images and the images from DECT scanner were compared quantitatively. Paired-sample t-tests were used for statistical analysis to show the consistency and precision of calculated HU values. Results: Excellent agreement was found between the DL-DECT and the ground truth DECT images (p values ranged from 0.50 to 0.95). Noise reduction up to 68% (from 163 HU to 51 HU) was achieved for DL-based VMCT imaging as compared to that obtained by using the standard DECT. For the DL-based VMCT, the maximum iodine contrast-to-noise ratio (CNR) for each patient (ranging from 15.1 to 16.6) was achieved at 40 keV. In addition to the enormous benefit of VMCT acquisition with merely a SECT image, an improvement of CNR as high as 55% (from 10.7 to 16.6) was attained with the proposed approach. Conclusions: This study demonstrates that high-quality VMCT images can be obtained with only a SECT scan.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

A Mean-Field Theory for Kernel Alignment with Random Features in Generative and Discriminative Models

We propose a novel supervised learning method to optimize the kernel in the maximum mean discrepancy generative adversarial networks (MMD GANs), and the kernel support vector machines (SVMs). Specifically, we characterize a distributionally robust optimization problem to compute a good distribution for the random feature model of Rahimi and Recht. Due to the fact that the distributional optimization is infinite dimensional, we consider a Monte-Carlo sample average approximation (SAA) to obtain a more tractable finite dimensional optimization problem. We subsequently leverage a particle stochastic gradient descent (SGD) method to solve the derived finite dimensional optimization problem. Based on a mean-field analysis, we then prove that the empirical distribution of the interactive particles system at each iteration of the SGD follows the path of the gradient descent flow on the Wasserstein manifold. We also establish the non-asymptotic consistency of the finite sample estimator. We evaluate our kernel learning method for the hypothesis testing problem by evaluating the kernel MMD statistics, and show that our learning method indeed attains better power of the test for larger threshold values compared to an untrained kernel. Moreover, our empirical evaluation on benchmark data-sets shows the advantage of our kernel learning approach compared to alternative kernel learning methods.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

A Mean-Field Theory for Learning the Schönberg Measure of Radial Basis Functions

We develop and analyze a projected particle Langevin optimization method to learn the distribution in the Schönberg integral representation of the radial basis functions from training samples. More specifically, we characterize a distributionally robust optimization method with respect to the Wasserstein distance to optimize the distribution in the Schönberg integral representation. To provide theoretical performance guarantees, we analyze the scaling limits of a projected particle online (stochastic) optimization method in the mean-field regime. In particular, we prove that in the scaling limits, the empirical measure of the Langevin particles converges to the law of a reflected Itô diffusion-drift process. Moreover, the drift is also a function of the law of the underlying process. Using Itô lemma for semi-martingales and Grisanov's change of measure for the Wiener processes, we then derive a Mckean-Vlasov type partial differential equation (PDE) with Robin boundary conditions that describes the evolution of the empirical measure of the projected Langevin particles in the mean-field regime. In addition, we establish the existence and uniqueness of the steady-state solutions of the derived PDE in the weak sense. We apply our learning approach to train radial kernels in the kernel locally sensitive hash (LSH) functions, where the training data-set is generated via a $k$-mean clustering method on a small subset of data-base. We subsequently apply our kernel LSH with a trained kernel for image retrieval task on MNIST data-set, and demonstrate the efficacy of our kernel learning approach. We also apply our kernel learning approach in conjunction with the kernel support vector machines (SVMs) for classification of benchmark data-sets.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Data-driven dose calculation algorithm based on deep learning

In this study we performed a feasibility investigation on implementing a fast and accurate dose calculation based on a deep learning technique. A two dimensional (2D) fluence map was first converted into a three dimensional (3D) volume using ray traversal algorithm. A 3D U-Net like deep residual network was then established to learn a mapping between this converted 3D volume, CT and 3D dose distribution. Therefore an indirect relationship was built between a fluence map and its corresponding 3D dose distribution without using significantly complex neural networks. 200 patients, including nasopharyngeal, lung, rectum and breast cancer cases, were collected and applied to train the proposed network. Additional 47 patients were randomly selected to evaluate the accuracy of the proposed method through comparing dose distributions, dose volume histograms (DVH) and clinical indices with the results from a treatment planning system (TPS), which was used as the ground truth in this study. Results: The proposed deep learning based dose calculation algorithm achieved good predictive performance. For 47 tested patients, the average per-voxel bias of the deep learning calculated value and standard deviation (normalized to the prescription), relative to the TPS calculation, is 0.17%. The average deep learning calculated values and standard deviations for relevant clinical indices were compared with the TPS calculated results and the t-test p-values demonstrated the consistency between them. Conclusions: In this study we developed a new deep learning based dose calculation method. This approach was evaluated by the clinical cases with different sites. Our results demonstrated its feasibility and reliability and indicated its great potential to improve the efficiency and accuracy of radiation dose calculation for different treatment modalities

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Deep Sinogram Completion with Image Prior for Metal Artifact Reduction in CT Images

Computed tomography (CT) has been widely used for medical diagnosis, assessment, and therapy planning and guidance. In reality, CT images may be affected adversely in the presence of metallic objects, which could lead to severe metal artifacts and influence clinical diagnosis or dose calculation in radiation therapy. In this paper, we propose a generalizable framework for metal artifact reduction (MAR) by simultaneously leveraging the advantages of image domain and sinogram domain-based MAR techniques. We formulate our framework as a sinogram completion problem and train a neural network (SinoNet) to restore the metal-affected projections. To improve the continuity of the completed projections at the boundary of metal trace and thus alleviate new artifacts in the reconstructed CT images, we train another neural network (PriorNet) to generate a good prior image to guide sinogram learning, and further design a novel residual sinogram learning strategy to effectively utilize the prior image information for better sinogram completion. The two networks are jointly trained in an end-to-end fashion with a differentiable forward projection (FP) operation so that the prior image generation and deep sinogram completion procedures can benefit from each other. Finally, the artifact-reduced CT images are reconstructed using the filtered backward projection (FBP) from the completed sinogram. Extensive experiments on simulated and real artifacts data demonstrate that our method produces superior artifact-reduced results while preserving the anatomical structures and outperforms other MAR methods.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Difficulty-aware Meta-learning for Rare Disease Diagnosis

Rare diseases have extremely low-data regimes, unlike common diseases with large amount of available labeled data. Hence, to train a neural network to classify rare diseases with a few per-class data samples is very challenging, and so far, catches very little attention. In this paper, we present a difficulty-aware meta-learning method to address rare disease classifications and demonstrate its capability to classify dermoscopy images. Our key approach is to first train and construct a meta-learning model from data of common diseases, then adapt the model to perform rare disease classification.To achieve this, we develop the difficulty-aware meta-learning method that dynamically monitors the importance of learning tasks during the meta-optimization stage. To evaluate our method, we use the recent ISIC 2018 skin lesion classification dataset, and show that with only five samples per class, our model can quickly adapt to classify unseen classes by a high AUC of 83.3%. Also, we evaluated several rare disease classification results in the public Dermofit Image Library to demonstrate the potential of our method for real clinical practice.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Dual-energy CT imaging from single-energy CT data with material decomposition convolutional neural network

Dual-energy computed tomography (DECT) is of great significance for clinical practice due to its huge potential to provide material-specific information. However, DECT scanners are usually more expensive than standard single-energy CT (SECT) scanners and thus are less accessible to undeveloped regions. In this paper, we show that the energy-domain correlation and anatomical consistency between standard DECT images can be harnessed by a deep learning model to provide high-performance DECT imaging from fully-sampled low-energy data together with single-view high-energy data, which can be obtained by using a scout-view high-energy image. We demonstrate the feasibility of the approach with contrast-enhanced DECT scans from 5,753 slices of images of twenty-two patients and show its superior performance on DECT applications. The deep learning-based approach could be useful to further significantly reduce the radiation dose of current premium DECT scanners and has the potential to simplify the hardware of DECT imaging systems and to enable DECT imaging using standard SECT scanners.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Multi-Domain Image Completion for Random Missing Input Data

Multi-domain data are widely leveraged in vision applications taking advantage of complementary information from different modalities, e.g., brain tumor segmentation from multi-parametric magnetic resonance imaging (MRI). However, due to possible data corruption and different imaging protocols, the availability of images for each domain could vary amongst multiple data sources in practice, which makes it challenging to build a universal model with a varied set of input data. To tackle this problem, we propose a general approach to complete the random missing domain(s) data in real applications. Specifically, we develop a novel multi-domain image completion method that utilizes a generative adversarial network (GAN) with a representational disentanglement scheme to extract shared skeleton encoding and separate flesh encoding across multiple domains. We further illustrate that the learned representation in multi-domain image completion could be leveraged for high-level tasks, e.g., segmentation, by introducing a unified framework consisting of image completion and segmentation with a shared content encoder. The experiments demonstrate consistent performance improvement on three datasets for brain tumor segmentation, prostate segmentation, and facial expression image completion respectively.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Noise2Context: Context-assisted Learning 3D Thin-layer Low Dose CT Without Clean Data

Computed tomography (CT) has played a vital role in medical diagnosis, assessment, and therapy planning, etc. In clinical practice, concerns about the increase of X-ray radiation exposure attract more and more attention. To lower the X-ray radiation, low-dose CT is often used in certain scenarios, while it will induce the degradation of CT image quality. In this paper, we proposed a training method that trained denoising neural networks without any paired clean data. we trained the denoising neural network to map one noise LDCT image to its two adjacent LDCT images in a singe 3D thin-layer low-dose CT scanning, simultaneously In other words, with some latent assumptions, we proposed an unsupervised loss function with the integration of the similarity between adjacent CT slices in 3D thin-layer lowdose CT to train the denoising neural network in an unsupervised manner. For 3D thin-slice CT scanning, the proposed virtual supervised loss function was equivalent to a supervised loss function with paired noisy and clean samples when the noise in the different slices from a single scan was uncorrelated and zero-mean. Further experiments on Mayo LDCT dataset and a realistic pig head were carried out and demonstrated superior performance over existing unsupervised methods.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Self-supervised Feature Learning via Exploiting Multi-modal Data for Retinal Disease Diagnosis

The automatic diagnosis of various retinal diseases from fundus images is important to support clinical decision-making. However, developing such automatic solutions is challenging due to the requirement of a large amount of human-annotated data. Recently, unsupervised/self-supervised feature learning techniques receive a lot of attention, as they do not need massive annotations. Most of the current self-supervised methods are analyzed with single imaging modality and there is no method currently utilize multi-modal images for better results. Considering that the diagnostics of various vitreoretinal diseases can greatly benefit from another imaging modality, e.g., FFA, this paper presents a novel self-supervised feature learning method by effectively exploiting multi-modal data for retinal disease diagnosis. To achieve this, we first synthesize the corresponding FFA modality and then formulate a patient feature-based softmax embedding objective. Our objective learns both modality-invariant features and patient-similarity features. Through this mechanism, the neural network captures the semantically shared information across different modalities and the apparent visual similarity between patients. We evaluate our method on two public benchmark datasets for retinal disease diagnosis. The experimental results demonstrate that our method clearly outperforms other self-supervised feature learning methods and is comparable to the supervised baseline.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Transformation Consistent Self-ensembling Model for Semi-supervised Medical Image Segmentation

Deep convolutional neural networks have achieved remarkable progress on a variety of medical image computing tasks. A common problem when applying supervised deep learning methods to medical images is the lack of labeled data, which is very expensive and time-consuming to be collected. In this paper, we present a novel semi-supervised method for medical image segmentation, where the network is optimized by the weighted combination of a common supervised loss for labeled inputs only and a regularization loss for both labeled and unlabeled data. To utilize the unlabeled data, our method encourages the consistent predictions of the network-in-training for the same input under different regularizations. Aiming for the semi-supervised segmentation problem, we enhance the effect of regularization for pixel-level predictions by introducing a transformation, including rotation and flipping, consistent scheme in our self-ensembling model. With the aim of semi-supervised segmentation tasks, we introduce a transformation consistent strategy in our self-ensembling model to enhance the regularization effect for pixel-level predictions. We have extensively validated the proposed semi-supervised method on three typical yet challenging medical image segmentation tasks: (i) skin lesion segmentation from dermoscopy images on International Skin Imaging Collaboration (ISIC) 2017 dataset, (ii) optic disc segmentation from fundus images on Retinal Fundus Glaucoma Challenge (REFUGE) dataset, and (iii) liver segmentation from volumetric CT scans on Liver Tumor Segmentation Challenge (LiTS) dataset. Compared to the state-of-the-arts, our proposed method shows superior segmentation performance on challenging 2D/3D medical images, demonstrating the effectiveness of our semi-supervised method for medical image segmentation.

0 citations0 reviewsNo code linkedOpen access