arXiv subjects

Ran Shi

Publications and source records attributed to Ran Shi.

6 recordsLinked to original sources

arXiv2026

A Clinician-Centered Pipeline for Annotation and Evaluation in Ultrasound AI Studies

Clinician-centered evaluation is critical for validating medical AI systems, especially in ultrasound imaging where quantitative metrics do not always capture clinical usability. Existing medical image platforms primarily focus on dataset labeling. They lack integrated support for blinded model comparison and reproducible evaluation workflows. We present a clinician-centered pipeline for remote annotation and evaluation in ultrasound AI studies. The proposed pipeline uses a centralized server and lightweight browser interfaces to enable clinicians to perform annotation, blinded ranking, and review without local dataset downloads. The pipeline also supports multi-rater participation, centralized result aggregation, and automated statistical analysis. We validate the pipeline in a fetal ultrasound segmentation study with six raters spanning expert, generalist, and non-expert experience levels. The system automatically generated Spearman correlation, Kendall's $\tau$, and top-1 selection statistics. Results indicated moderate to strong agreement across experts and other groups. The blinded evaluation results showed a tendency for later active learning models to be preferred. These outcomes suggest that the pipeline can support clinician-centered annotation and reproducible human-\ac{AI} evaluation studies in ultrasound imaging. The proposed pipeline is available on \href{https://github.com/13204942/SonoRate}{GitHub}.

cs.HC↗

arXiv2024

LDPM: Towards undersampled MRI reconstruction with MR-VAE and Latent Diffusion Prior

Diffusion models, as powerful generative models, have found a wide range of applications and shown great potential in solving image reconstruction problems. Some works attempted to solve MRI reconstruction with diffusion models, but these methods operate directly in pixel space, leading to higher computational costs for optimization and inference. Latent diffusion models, pre-trained on natural images with rich visual priors, are expected to solve the high computational cost problem in MRI reconstruction by operating in a lower-dimensional latent space. However, direct application to MRI reconstruction faces three key challenges: (1) absence of explicit control mechanisms for medical fidelity, (2) domain gap between natural images and MR physics, and (3) undefined data consistency in latent space. To address these challenges, a novel Latent Diffusion Prior-based undersampled MRI reconstruction (LDPM) method is proposed. Our LDPM framework addresses these challenges by: (1) a sketch-guided pipeline with a two-step reconstruction strategy, which balances perceptual quality and anatomical fidelity, (2) an MRI-optimized VAE (MR-VAE), which achieves an improvement of approximately 3.92 dB in PSNR for undersampled MRI reconstruction compared to that with SD-VAE \cite{sd}, and (3) Dual-Stage Sampler, a modified version of spaced DDPM sampler, which enforces high-fidelity reconstruction in the latent space. Experiments on the fastMRI dataset\cite{fastmri} demonstrate the state-of-the-art performance of the proposed method and its robustness across various scenarios. The effectiveness of each module is also verified through ablation experiments.

eess.IV↗

arXiv2023

Substrate-Selective Adhesion of Metal Nanoparticles to Graphene Devices

Nanostructured electronic devices, such as those based on graphene, are typically grown on top of the insulator SiO2. Their exposure to a flux of small size-selected silver nanoparticles has revealed remarkably selective adhesion: the graphene channel can be made fully metallized while the insulating substrate remains coverage-free. This conspicuous contrast derives from the low binding energy between the metal nanoparticles and a contaminant-free passivated silica surface. In addition to providing physical insight into nanoparticle adhesion, this effect may be of value in applications involving deposition of metallic layers on device working surfaces: it eliminates the need for masking the insulating region and the associated extensive and potentially deleterious pre- and postprocessing.

physics.atm-clus↗

arXiv2021

Personal Fixations-Based Object Segmentation with Object Localization and Boundary Preservation

As a natural way for human-computer interaction, fixation provides a promising solution for interactive image segmentation. In this paper, we focus on Personal Fixations-based Object Segmentation (PFOS) to address issues in previous studies, such as the lack of appropriate dataset and the ambiguity in fixations-based interaction. In particular, we first construct a new PFOS dataset by carefully collecting pixel-level binary annotation data over an existing fixation prediction dataset, such dataset is expected to greatly facilitate the study along the line. Then, considering characteristics of personal fixations, we propose a novel network based on Object Localization and Boundary Preservation (OLBP) to segment the gazed objects. Specifically, the OLBP network utilizes an Object Localization Module (OLM) to analyze personal fixations and locates the gazed objects based on the interpretation. Then, a Boundary Preservation Module (BPM) is designed to introduce additional boundary information to guard the completeness of the gazed objects. Moreover, OLBP is organized in the mixed bottom-up and top-down manner with multiple types of deep supervision. Extensive experiments on the constructed PFOS dataset show the superiority of the proposed OLBP network over 17 state-of-the-art methods, and demonstrate the effectiveness of the proposed OLM and BPM components. The constructed PFOS dataset and the proposed OLBP network are available at https://github.com/MathLee/OLBPNet4PFOS.

cs.CV↗

arXiv2015

Modeling Covariate Effects in Group Independent Component Analysis with Applications to Functional Magnetic Resonance Imaging

Independent component analysis (ICA) is a powerful computational tool for separating independent source signals from their linear mixtures. ICA has been widely applied in neuroimaging studies to identify and characterize underlying brain functional networks. An important goal in such studies is to assess the effects of subjects' clinical and demographic covariates on the spatial distributions of the functional networks. Currently, covariate effects are not incorporated in existing group ICA decomposition methods. Hence, they can only be evaluated through ad-hoc approaches which may not be accurate in many cases. In this paper, we propose a hierarchical covariate ICA model that provides a formal statistical framework for estimating and testing covariate effects in ICA decomposition. A maximum likelihood method is proposed for estimating the covariate ICA model. We develop two expectation-maximization (EM) algorithms to obtain maximum likelihood estimates. The first is an exact EM algorithm, which has analytically tractable E-step and M-step. Additionally, we propose a subspace-based approximate EM, which can significantly reduce computational time while still retain high model-fitting accuracy. Furthermore, to test covariate effects on the functional networks, we develop a voxel-wise approximate inference procedure which eliminates the needs of computationally expensive covariance estimation. The performance of the proposed methods is evaluated via simulation studies. The application is illustrated through an fMRI study of Zen meditation.

stat.AP↗

arXiv2015

Thresholded Multiscale Gaussian Processes with Application to Bayesian Feature Selection for Massive Neuroimaging Data

Motivated by the needs of selecting important features for massive neuroimaging data, we propose a spatially varying coefficient model (SVCMs) with sparsity and piecewise smoothness imposed on the coefficient functions. A new class of nonparametric priors is developed based on thresholded multiresolution Gaussian processes (TMGP). We show that the TMGP has a large support on a space of sparse and piecewise smooth functions, leading to posterior consistency in coefficient function estimation and feature selection. Also, we develop a method for prior specifications of thresholding parameters in TMGPs and discuss their theoretical properties. Efficient posterior computation algorithms are developed by adopting a kernel convolution approach, where a modified square exponential kernel is chosen taking the advantage that the analytical form of the eigen decomposition is available. Based on simulation studies, we demonstrate that our methods can achieve better performance in estimating the spatially varying coefficient. Also, the proposed model has been applied to an analysis of resting state functional magnetic resonance imaging (Rs-fMRI) data from the Autism Brain Imaging Data Exchange (ABIDE) study, it provides biologically meaningful results.

stat.ME↗