A5100431313- Image and Signal Denoising Methods
- Human Pose and Action Recognition
- Advanced Image Fusion Techniques
- Video Surveillance and Tracking Methods
- Advanced Adaptive Filtering Techniques
- Anomaly Detection Techniques and Applications
- 3D Shape Modeling and Analysis
- Advanced Vision and Imaging
- Remote-Sensing Image Classification
- Advanced Image and Video Retrieval Techniques
- Advanced Neural Network Applications
- Digital Filter Design and Implementation
- Image Retrieval and Classification Techniques
- Advanced SAR Imaging Techniques
- 3D Surveying and Cultural Heritage
- Computer Graphics and Visualization Techniques
- Radar Systems and Signal Processing
- Additive Manufacturing Materials and Processes
- Speech and Audio Processing
- High Entropy Alloys Studies
- Sparse and Compressive Sensing Techniques
- Image and Object Detection Techniques
- Air Quality and Health Impacts
- Music and Audio Processing
- Robotics and Sensor-Based Localization
Shanghai University
2016-2025
Sun Yat-sen University Cancer Center
2025
Institute of Earth Environment
2024-2025
Chinese Academy of Sciences
2019-2025
Fujian Medical University
2024-2025
Union Hospital
2025
University of Sheffield
2025
Capital Medical University
2024-2025
Westlake University
2024-2025
Sun Yat-sen University
2025
Users' locations are important to many applications such as targeted advertisement and news recommendation. In this paper, we focus on the problem of profiling users' home in context social network (Twitter). The is nontrivial, because signals, which may help identify a user's location, scarce noisy. We propose unified discriminative influence model, named UDI, solve problem. To overcome challenge UDI integrates signals observed from both (friends) user-centric data (tweets) probabilistic...
In this paper we present an extension of Direct Sparse Odometry (DSO) [1] to a monocular visual SLAM system with loop closure detection and pose-graph optimization (LDSO). As direct technique, DSO can utilize any image pixel sufficient intensity gradient, which makes it robust even in featureless areas. LDSO retains robustness, while at the same time ensuring repeatability some these points by favoring corner features tracking frontend. This allows reliably detect candidates conventional...
In this paper we investigate the problem of inducing a distribution over three-dimensional structures given two-dimensional views multiple objects taken from unknown viewpoints. Our approach called "projective generative adversarial networks" (PrGANs) trains deep model 3D shapes whose projections match distributions input 2D views. The addition projection module allows us to infer underlying shape without using any 3D, viewpoint information, or annotation during learning phase. We show that...
This paper studies the BERT pretraining of video transformers. It is a straightforward but worth-studying extension given recent success from image We introduce BEVT which decouples representation learning into spatial and temporal dynamics learning. In particular, first performs masked modeling on data, then conducts jointly with data. design motivated by two observations: 1) transformers learned datasets provide decent priors that can ease transformers, are often times...
Extreme temperature events (ETEs), including heat wave and cold spell, have been linked to myocardial infarction (MI) morbidity; however, their effects on MI mortality are less clear. Although ambient fine particulate matter (PM
Benefiting from masked visual modeling, self-supervised video representation learning has achieved remarkable progress. However, existing methods focus on representations scratch through reconstructing low-level features like raw pixel values. In this paper, we propose distillation (MVD), a simple yet effective two-stage feature modeling framework for learning: firstly pretrain an image (or video) model by recovering of patches, then use the resulting as targets modeling. For choice teacher...
<title>Abstract</title> With technological advancements, we can now capture rich dialogue content, tones, textual information, and visual data through tools like microphones, the internet, cameras. However, relying solely on a single modality for emotion analysis often fails to accurately reflect true emotional state, as this approach overlooks dynamic correlations between different modalities. To address this, our study introduces multimodal recognition method that combines tensor...
Nickel-based superalloys have been widely used in the aerospace, petrochemical, and marine fields others because of their good oxidation resistance, corrosion stability, reliability at various temperatures. However, as a nickel-based superalloy is kind processed material, cutting process large amount heat generated due to interaction between tool workpiece. At same time, low thermal conductivity workpiece causes accumulate contact point, resulting serious wear, reduced life, frequent...
Training text-to-image models with web scale image-text pairs enables the generation of a wide range visual concepts from text. However, these pre-trained often face challenges when it comes to generating highly aesthetic images. This creates need for alignment post pre-training. In this paper, we propose quality-tuning effectively guide model exclusively generate visually appealing images, while maintaining generality across concepts. Our key insight is that supervised fine-tuning set...
Abstract Metal halide perovskites have demonstrated considerable promise across various optoelectronic applications. Surface passivation serves as a pivotal strategy to obtain high‐quality perovskite materials, either in manner of bulk thin film or nanocrystal, with superior properties and stability. The current research focus this regard primarily revolves around the use organic molecules passivate surface perovskites. However, always suffer from chemical instability weak secondary bonding...
Sampling is a core component for many graphics applications including rendering, imaging, animation, and geometry processing. The efficacy of these often crucially depends upon the distribution quality underlying samples. While uniform sampling can be analyzed by using existing spatial spectral methods, cannot easily extended to general non-uniform settings, such as adaptive, anisotropic, or non-Euclidean domains. We present new methods analyzing sample distributions. Our key insight that...
Point samples with different spectral noise properties (often defined using color names such as white, blue, green, and red) are important for many science engineering disciplines including computer graphics. While existing techniques can easily produce white blue samples, relatively little is known generating other patterns. In particular, no single algorithm available to generate patterns according user-defined spectra. this paper, we describe an point that match a Fourier spectrum...
Named entity recognition (NER) is an essential part of natural language processing tasks. Chinese NER task different from the many European languages due to lack delimiters. Therefore, Word Segmentation (CWS) usually regarded as first step NER. However, word-based models relying on CWS are more vulnerable incorrectly segmented boundaries and presence out-of-vocabulary (OOV) words. In this paper, we propose a novel character-based Gated Convolutional Recurrent neural network with Attention...