A5068526974- Data Visualization and Analytics
- Data Management and Algorithms
- Computer Graphics and Visualization Techniques
- Image Retrieval and Classification Techniques
- Advanced Vision and Imaging
- Neural dynamics and brain function
- Parasitic infections in humans and animals
- Remote Sensing and LiDAR Applications
- 3D Shape Modeling and Analysis
- Advanced Image and Video Retrieval Techniques
- Congenital Anomalies and Fetal Surgery
- Geoscience and Mining Technology
- Gastric Cancer Management and Outcomes
- stochastic dynamics and bifurcation
- Time Series Analysis and Forecasting
- Topological and Geometric Data Analysis
- Complex Network Analysis Techniques
- Metastasis and carcinoma case studies
- Advanced Neural Network Applications
- Medical Image Segmentation Techniques
- Photoreceptor and optogenetics research
- Video Analysis and Summarization
- Advanced Clustering Algorithms Research
- Cell Image Analysis Techniques
- Image and Video Quality Assessment
Kunming University
2024-2025
Renmin University of China
2024-2025
Shanghai University of Electric Power
2023-2024
University of Hong Kong
2024
Shandong University
2016-2023
Shandong University of Science and Technology
2019-2023
Electric Power University
2023
Shanghai University
2023
Tianjin University of Traditional Chinese Medicine
2023
California State University, Sacramento
2023
Unsupervised co-analysis of a set shapes is difficult problem since the geometry alone cannot always fully describe semantics shape parts. In this paper, we propose semi-supervised learning method where user actively assists in by iteratively providing inputs that progressively constrain system. We introduce novel constrained clustering based on spring system which embeds elements to better respect their inter-distances feature space together with user-given constraints. also present an...
We propose a new method to tackle the mapping challenge from time-series data spatial image in field of seismic exploration, i.e., reconstructing velocity model directly by deep neural networks (DNNs). The conventional way addressing this ill-posed inversion problem is through iterative algorithms, which suffer poor nonlinear and strong nonuniqueness. Other attempts may either import human intervention errors or underuse data. for DNNs mainly lies weak correspondence, uncertain...
To determine the true community prevalence of human cystic (CE) and alveolar (AE) echinococcosis (hydatid disease) in a highly endemic region Ningxia Hui, China, by detecting asymptomatic cases.Using hospital records "AE-risk" landscape patterns we selected study communities predicted to be at risk Guyuan, Longde Xiji counties. We conducted surveys 4773 individuals from 26 villages 2002 2003 using questionnaire analysis, ultrasound examination serology.Ultrasound serology showed range...
Fuzzy clustering assigns a probability of membership for datum to cluster, which veritably reflects real-world scenarios but significantly increases the complexity understanding fuzzy clusters. Many studies have demonstrated that visualization techniques multi-dimensional data are beneficial understand However, no empirical evidence exists on effectiveness and efficiency these in solving analytical tasks featured by In this paper, we conduct controlled experiment evaluate ability clusters...
Line graphs are usually considered to be the best choice for visualizing time series data, whereas sometimes also scatter plots used showing main trends. So far there no guidelines that indicate which of these visualization methods better display trends in a given canvas. Assuming information is its overall trend, we propose an algorithm automatically picks method reveals this trend best. This achieved by measuring visual consistency between curve represented LOESS fit and described plot or...
Dimensionality reduction (DR) is a common strategy for visual analysis of labeled high-dimensional data. Low-dimensional representations the data help, instance, to explore class separability and spatial distribution Widely-used unsupervised DR methods like PCA do not aim maximize separation, while supervised LDA often assume certain distributions take perceptual capabilities humans into account. These issues make them ineffective complicated structures. Towards filling this gap, we present...
Appropriate choice of colors significantly aids viewers in understanding the structures multiclass scatterplots and becomes more important with a growing number data points groups. An appropriate color mapping is also an parameter for creation aesthetically pleasing scatterplot. Currently, users visualization software routinely rely on mappings that have been pre-defined by software. A default mapping, however, cannot ensure optimal perceptual separability between groups, sometimes may even...
The multidimensional transfer function is a flexible and effective tool for exploring volume data. However, designing an appropriate trial-and-error process remains challenge. In this paper, we propose novel exploration scheme that explores volumetric structures in the feature space by modeling using Gaussian mixture model (GMM). Our new approach has three distinctive advantages. First, initial separation can be automatically achieved through GMM estimation. Second, calculated Gaussians...
We introduce projective analysis for semantic segmentation and labeling of 3D shapes. The treats an input shape as a collection 2D projections, labels each projection by transferring knowledge from existing labeled images, back-projects fuses the labelings on shape. image-space involves matching projected binary images objects based novel bi-class Hausdorff distance . is topology-aware accounting internal holes in figures it applied to piecewise-linearly warped object projections compensate...
We present an improved stress majorization method that incorporates various constraints, including directional constraints without the necessity of solving a constraint optimization problem. This is achieved by reformulating function to impose on both edge vectors and lengths instead just (node distances). unified framework for constrained unconstrained graph visualizations, where we can model most existing layout as well develop new ones such star shapes cluster separation within...
Since multiple criteria can be adopted to estimate the similarity among given data points, problem regarding diverse representations of pairwise relations is brought about. To address this issue, a novel self-adaptive multimeasure (SAMM) fusion proposed, such that different measure functions adaptively merged into unified measure. Different from other approaches, we optimize as variable instead presetting it priori, evaluated based on integrating various measures. further obtain associated...
Dimensionality Reduction (DR) techniques can generate 2D projections and enable visual exploration of cluster structures high-dimensional datasets. However, different DR would yield various patterns, which significantly affect the performance analysis tasks. We present results a user study that investigates influence on analysis. Our focuses most concerned property types, namely linearity locality, evaluates twelve representative cover properties. Four controlled experiments were conducted...
Time-series data-usually presented in the form of lines-plays an important role many domains such as finance, meteorology, health, and urban informatics. Yet, little has been done to support interactive exploration large-scale time-series data, which requires a clutter-free visual representation with low-latency interactions. In this paper, we contribute novel line-segment-based KD-tree method enable analysis time series. Our enables not only fast queries over series selected regions...
Graph convolution network-based approaches have been recently used to model region-wise relationships in region-level prediction problems urban computing. Each relationship represents a kind of spatial dependency, such as distance or functional similarity. To incorporate multiple into feature extraction, we define the problem multi-modal machine learning on multi-graph networks. Leveraging advantage learning, propose develop modality interaction mechanisms for this order reduce...
We present EdWordle, a method for consistently editing word clouds. At its heart, EdWordle allows users to move and edit words while preserving the neighborhoods of other words. To do so, we combine constrained rigid body simulation with neighborhood-aware local Wordle algorithm update cloud create very compact layouts. The consistent stable behavior enables new forms clouds such as storytelling in which position is carefully edited. compare our approach state-of-the-art methods show that...
Abstract Some problems such as the decline of new labor force, increase retired force emerge because complex and changeable market environment, consequently exacerbating staffing problem in retail industry. Also, unreasonable distribution personnel, there are few people busy hours too many idle hours, which causes waste labor. To address this issue, we analyzed time series sales volume industry detail, processed data with feature engineering for predicting in-store future. At same time,...