BZPEERPapers, people, evidence and research paths
Menu

Discover

SearchFind papers, people, topics, institutions and opportunities from one place.GraphSee how papers, authors, topics and evidence connect.

Research tools

HomeContinue from the papers, people and topics that matter now.CollectionsTurn saved papers into reading lists, evidence maps and shareable context.ResearchReconstruct the path that led you from search to paper to question.CollaborationMove from a paper, topic or expert request into a focused thread.PublishCreate full-text research articles with metadata, formulas and versions.

Network

ExpertsFind specialists whose work and reviews explain why they can help.CommunitiesJoin research groups organized around papers, topics and shared evidence.

Opportunities

ListingsBrowse roles, calls and collaborations connected to real research context.

Account

Log inReturn to your saved papers, graph views and active threads.Create accountStart saving papers, building a research trail and joining teams.
Log inCreate account

Source author record

Qiao Liu

Qiao Liu appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record
math.APComputer VisionArtificial IntelligenceMachine Learningeess.IVMultimedia

Catalog footprint

What is connected

15works
6topics
4close collaborators

Actions

Connect this record

Log in to claim
Open graphBrowse works

Research graph

See the researcher in context

Open full explorer

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

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

Published work

15 published item(s)

preprint2026arXiv

Missingness-aware Data Imputation via AI-powered Bayesian Generative Modeling

Missing data imputation remains a fundamental challenge in modern data science, especially when uncertainty quantification is essential. In this work, we propose MissBGM, an AI-powered missing data imputation method via Bayesian generative modeling that bridges the expressive flexibility of neural networks with the statistical rigor of Bayesian inference. Unlike existing methods that often focus on point estimates or treat the missingness mechanism implicitly, MissBGM explicitly and jointly models the data-generating and missingness mechanisms, providing principled posterior uncertainty over imputations rather than a single point estimate. We develop a stochastic optimization framework with alternating updates among missing values, model parameters, and latent variables until convergence. Our theoretical analysis shows that estimates of missing values from MissBGM converge consistently under mild assumptions. Empirically, we demonstrate that MissBGM achieves superior performance over traditional imputers and recent neural network-based methods across extensive experimental settings. These results establish MissBGM as a principled and scalable solution for modern missing data imputation. The code for MissBGM is open sourced at https://github.com/liuq-lab/MissBGM.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Accurate Bounding-box Regression with Distance-IoU Loss for Visual Tracking

Most existing trackers are based on using a classifier and multi-scale estimation to estimate the target state. Consequently, and as expected, trackers have become more stable while tracking accuracy has stagnated. While trackers adopt a maximum overlap method based on an intersection-over-union (IoU) loss to mitigate this problem, there are defects in the IoU loss itself, that make it impossible to continue to optimize the objective function when a given bounding box is completely contained within/without another bounding box; this makes it very challenging to accurately estimate the target state. Accordingly, in this paper, we address the above-mentioned problem by proposing a novel tracking method based on a distance-IoU (DIoU) loss, such that the proposed tracker consists of target estimation and target classification. The target estimation part is trained to predict the DIoU score between the target ground-truth bounding-box and the estimated bounding-box. The DIoU loss can maintain the advantage provided by the IoU loss while minimizing the distance between the center points of two bounding boxes, thereby making the target estimation more accurate. Moreover, we introduce a classification part that is trained online and optimized with a Conjugate-Gradient-based strategy to guarantee real-time tracking speed. Comprehensive experimental results demonstrate that the proposed method achieves competitive tracking accuracy when compared to state-of-the-art trackers while with a real-time tracking speed.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Active Learning for Deep Visual Tracking

Convolutional neural networks (CNNs) have been successfully applied to the single target tracking task in recent years. Generally, training a deep CNN model requires numerous labeled training samples, and the number and quality of these samples directly affect the representational capability of the trained model. However, this approach is restrictive in practice, because manually labeling such a large number of training samples is time-consuming and prohibitively expensive. In this paper, we propose an active learning method for deep visual tracking, which selects and annotates the unlabeled samples to train the deep CNNs model. Under the guidance of active learning, the tracker based on the trained deep CNNs model can achieve competitive tracking performance while reducing the labeling cost. More specifically, to ensure the diversity of selected samples, we propose an active learning method based on multi-frame collaboration to select those training samples that should be and need to be annotated. Meanwhile, considering the representativeness of these selected samples, we adopt a nearest neighbor discrimination method based on the average nearest neighbor distance to screen isolated samples and low-quality samples. Therefore, the training samples subset selected based on our method requires only a given budget to maintain the diversity and representativeness of the entire sample set. Furthermore, we adopt a Tversky loss to improve the bounding box estimation of our tracker, which can ensure that the tracker achieves more accurate target states. Extensive experimental results confirm that our active learning-based tracker (ALT) achieves competitive tracking accuracy and speed compared with state-of-the-art trackers on the seven most challenging evaluation benchmarks.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Graph Convolutional Networks for Multi-modality Medical Imaging: Methods, Architectures, and Clinical Applications

Image-based characterization and disease understanding involve integrative analysis of morphological, spatial, and topological information across biological scales. The development of graph convolutional networks (GCNs) has created the opportunity to address this information complexity via graph-driven architectures, since GCNs can perform feature aggregation, interaction, and reasoning with remarkable flexibility and efficiency. These GCNs capabilities have spawned a new wave of research in medical imaging analysis with the overarching goal of improving quantitative disease understanding, monitoring, and diagnosis. Yet daunting challenges remain for designing the important image-to-graph transformation for multi-modality medical imaging and gaining insights into model interpretation and enhanced clinical decision support. In this review, we present recent GCNs developments in the context of medical image analysis including imaging data from radiology and histopathology. We discuss the fast-growing use of graph network architectures in medical image analysis to improve disease diagnosis and patient outcomes in clinical practice. To foster cross-disciplinary research, we present GCNs technical advancements, emerging medical applications, identify common challenges in the use of image-based GCNs and their extensions in model interpretation, large-scale benchmarks that promise to transform the scope of medical image studies and related graph-driven medical research.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

LSOTB-TIR:A Large-Scale High-Diversity Thermal Infrared Object Tracking Benchmark

In this paper, we present a Large-Scale and high-diversity general Thermal InfraRed (TIR) Object Tracking Benchmark, called LSOTBTIR, which consists of an evaluation dataset and a training dataset with a total of 1,400 TIR sequences and more than 600K frames. We annotate the bounding box of objects in every frame of all sequences and generate over 730K bounding boxes in total. To the best of our knowledge, LSOTB-TIR is the largest and most diverse TIR object tracking benchmark to date. To evaluate a tracker on different attributes, we define 4 scenario attributes and 12 challenge attributes in the evaluation dataset. By releasing LSOTB-TIR, we encourage the community to develop deep learning based TIR trackers and evaluate them fairly and comprehensively. We evaluate and analyze more than 30 trackers on LSOTB-TIR to provide a series of baselines, and the results show that deep trackers achieve promising performance. Furthermore, we re-train several representative deep trackers on LSOTB-TIR, and their results demonstrate that the proposed training dataset significantly improves the performance of deep TIR trackers. Codes and dataset are available at https://github.com/QiaoLiuHit/LSOTB-TIR.

0 citations0 reviewsNo code linkedOpen access
preprint2015arXiv

Gevrey analyticity of solutions to the 3D nematic liquid crystal flows in critical Besov space

We show that the solution to the Cauchy problem of the 3D nematic liquid crystal flows, with initial data belongs to a critical Besov space, belongs to a Gevrey class. More precisely, it is proved that for any $(u_{0},d_{0} - \overline{d}_{0}) \in \dot{B}^{\frac{3}{p}-1}_{p,1} (\mathbb{R}^{3}) \times \dot{B}^{\frac{3}{q}}_{q,1} (\mathbb{R}^{3})$ with some suitable conditions imposed on $p, q\in(1,\infty)$, there exists $T^{*}>0$ depending only on initial data, such that the nematic liquid crystal flows admits a unique solution $(u,d)$ on $\mathbb{R}^{3} \times (0,T^{*})$, and satisfies \begin{align*} \|e^{\sqrt{t} Λ_{1}}{u}(t) \|_{\widetilde{L}^{\infty}_{T^{*}} (\dot{B}^{\frac{3}{p}-1}_{p,1}) \cap \widetilde{L}^{1}_{T^{*}} (\dot{B}^{\frac{3}{p}+1}_{p,1})} + \|e^{\sqrt{t}Λ_{1}} ({d}(t)- \overline{d}_{0}) \|_{\widetilde{L}^{\infty}_{T^{*}} (\dot{B}^{\frac{3}{q}}_{q,1}) \cap \widetilde{L}^{1}_{T^{*}}(\dot{B}^{\frac{3}{q}+2}_{q,1})} < \infty. \end{align*} Here, $\overline{d}_{0} \in \mathbb{S}^{2}$ is a constant unit vector, and $Λ_{1}$ is the Fourier multiplier whose symbol is given by $|ξ|_{1}=|ξ_{1}|+|ξ_{2}|+|ξ_{3}|$. Moreover, if the initial data is sufficiently small enough, then $T^{*}=\infty$. As a consequence of the method, decay estimates of higher-order derivatives of solutions in Besov spaces are deduced.

0 citations0 reviewsNo code linkedOpen access
preprint2015arXiv

Global well-posedness of the 2D nonhomogeneous incompressible nematic liquid crystal flows with vacuum

This paper concerns the Cauchy problem of the two-dimensional (2D) nonhomogeneous incompressible nematic liquid crystal flows on the whole space $\mathbb{R}^{2}$ with vacuum as far field density. It is proved that the 2D nonhomogeneous incompressible nematic liquid crystal flows admits a unique global strong solution provided the initial data density and the gradient of orientation decay not too slow at infinity, and the initial orientation satisfies a geometric condition (see \eqref{eq1.3}). In particular, the initial data can be arbitrarily large and the initial density may contain vacuum states and even have compact support. As a byproduct, the large time behavior of the solution is also obtained.

0 citations0 reviewsNo code linkedOpen access
preprint2015arXiv

Well-posedness and decay for the dissipative system modeling electro-hydrodynamics in negative Besov spaces

In \cite{GW12} (Y. Guo, Y. Wang, Decay of dissipative equations and negative Sobolev spaces, Commun. Partial Differ. Equ. 37 (2012) 2165--2208), Y. Guo and Y. Wang developed a general new energy method for proving the optimal time decay rates of the solutions to dissipative equations. In this paper, we generalize this method in the framework of homogeneous Besov spaces. Moreover, we apply this method to a model arising from electro-hydrodynamics, which is a coupled system of the Navier-Stokes equations and the Poisson-Nernst-Planck equations through charge transport and external forcing terms. We show that the negative Besov norms are preserved along time evolution, and obtain the optimal temporal decay rates of the higher-order spatial derivatives of solutions by the Fourier splitting approach and the interpolation techniques.

0 citations0 reviewsNo code linkedOpen access
preprint2014arXiv

On the temporal decay of solutions to the two-dimensional nematic liquid crystal flows

We consider the temporal decay estimates for weak solutions to the two-dimensional nematic liquid crystal flows, and we show that the energy norm of a global weak solution has non-uniform decay \begin{align*} \|u(t)\|_{L^{2}}+\|\nabla d(t)\|_{L^{2}}\rightarrow 0\quad \text{ as } t\rightarrow \infty, \end{align*} under suitable conditions on the initial data. We also show the exact rate of the decay (uniform decay) of the energy norm of the global weak solution.

0 citations0 reviewsNo code linkedOpen access
preprint2013arXiv

Logarithmical Blow-up Criteria for the Nematic Liquid Crystal Flows

We investigate the blow-up criterion for the local in time classical solution of the nematic liquid crystal flows in dimension two and three. More precisely, $0<T_{*}<+\infty$ is the maximal time interval if and only if (i) for $n=3$, {align*} \int_{0}^{T_{*}}\frac{\|ω\|_{\dot{B}^{0}_{\infty,\infty}}+\|\nabla d\|_{\dot{B}^{0}_{\infty,\infty}}^{2}}{\sqrt{1+\text{ln}(e+\|ω\|_{\dot{B}^{0}_{\infty,\infty}} +\|\nabla d\|_{\dot{B}^{0}_{\infty,\infty}})}}\text{d}t=\infty, {align*} or {align*} \int_{0}^{T_{*}}\frac{\|\nabla u\|_{\dot{B}^{-1}_{\infty,\infty}}^{2}+\|\nabla d\|_{\dot{B}^{0}_{\infty,\infty}}^{2}}{\sqrt{1+\text{ln}(e+\|\nabla u\|_{\dot{B}^{-1}_{\infty,\infty}} +\|\nabla d\|_{\dot{B}^{0}_{\infty,\infty}})}}\text{d}t=\infty; {align*} and (ii) for $n=2$, {align*} \int_{0}^{T_{*}}\frac{\|\nabla d\|_{\dot{B}^{0}_{\infty,\infty}}^{2}}{\sqrt{1+\text{ln}(e +\|\nabla d\|_{\dot{B}^{0}_{\infty,\infty}})}}\text{d}t=\infty. {align*}

0 citations0 reviewsNo code linkedOpen access
preprint2013arXiv

Logarithmically Improved Blow-up Criteria for a Phase Field Navier-Stokes Vesicle-Fluid Interaction Model

In this paper, we study a hydrodynamical system modeling the deformation of vesicle membrane under external incompressible viscous flow fields. The system is in the Eulerian formulation and is governed by the coupling of the incompressible Navier-Stokes equations with a phase field equation. In the three dimensional case, we establish two logarithmically improved blow-up criteria for local smooth solutions of this system in terms of the vorticity field only in the homogeneous Besov spaces.

0 citations0 reviewsNo code linkedOpen access
preprint2012arXiv

A regularity criterion for the solution of the nematic liquid crystal flows in terms of $\dot{B}^{-1}_{\infty,\infty}$-norm

In this paper, we investigate regularity criterion for the solution of the nematic liquid crystal flows in dimension three and two. We prove the solution $(u,d)$ is smooth up to time $T$ provided that there exists a positive constant $\varepsilon_{0}>0$ such that (i) for n=3, |(u,\nabla d)|_{L^{\infty}(0,T;\dot{B}^{-1}_{\infty,\infty})}\leq \varepsilon_{0}, and (ii) for $n=2$, |\nabla d|_{L^{\infty}(0,T;\dot{B}^{-1}_{\infty,\infty})}\leq \varepsilon_{0}.

0 citations0 reviewsNo code linkedOpen access
preprint2011arXiv

A Beale--Kato--Majda criterion for the 3-D Compressible Nematic Liquid Crystal Flows with Vacuum

In this paper, we prove a Beale--Kato--Majda blow-up criterion in terms of the gradient of the velocity only for the strong solution to the 3-D compressible nematic liquid crystal flows with nonnegative initial densities. More precisely, the strong solution exists globally if the $L^{1}(0,T;L^{\infty})$-norm of the gradient of the velocity $u$ is bounded. Our criterion improves the recent result of X. Liu and L. Liu (\cite{LL}, A blow-up criterion for the compressible liquid crystals system, arXiv:1011.4399v2 [math-ph] 23 Nov. 2010).

0 citations0 reviewsNo code linkedOpen access
preprint2011arXiv

Existence of Solutions for the Debye-Hückel System with Low Regularity Initial Data

In this paper we study existence of solutions for the Cauchy problem of the Debye-Hückel system with low regularity initial data. By using the Chemin-Lerner time-space estimate for the heat equation, we prove that there exists a unique local solution if the initial data belongs to the Besov space $\dot{B}^{s}_{p,q}(\mathbb{R}^{n})$ for $-3/2<s\leq-2+\frac{n}{2}$, $p=\frac{n}{s+2}$ and $1\leq q\leq \infty$, and furthermore, if the initial data is sufficiently small then the solution is global. This result improves the regularity index of the initial data space in previous results on this model.

0 citations0 reviewsNo code linkedOpen access
preprint2011arXiv

Global Existence and Stability for a Hydrodynamic System in the Nematic Liquid Crystal Flows

In this paper we consider a coupled hydrodynamical system which involves the Navier-Stokes equations for the velocity field and kinematic transport equations for the molecular orientation field. By applying the Chemin-Lerner's time-space estimates for the heat equation and the Fourier localization technique, we prove that when initial data belongs to the critical Besov spaces with negative-order, there exists a unique local solution, and this solution is global when initial data is small enough. As a corollary, we obtain existence of global self-similar solutions. In order to figure out the relation between the solution obtained here and weak solution of standard sense, we establish a stability result, which yields in a direct way that all global weak solutions associated with the same initial data must coincide with the solution obtained here, namely, weak-strong uniqueness holds.

0 citations0 reviewsNo code linkedOpen access