A5100449791- Advanced Vision and Imaging
- Video Coding and Compression Technologies
- Advanced Image Processing Techniques
- Advanced Data Compression Techniques
- 3D Shape Modeling and Analysis
- Image and Video Quality Assessment
- Computer Graphics and Visualization Techniques
- Remote Sensing and LiDAR Applications
- 3D Surveying and Cultural Heritage
- Image Enhancement Techniques
- Image and Signal Denoising Methods
- Generative Adversarial Networks and Image Synthesis
- Advanced Image and Video Retrieval Techniques
- Advanced Neural Network Applications
- Multimedia Communication and Technology
- Image Processing Techniques and Applications
- Visual Attention and Saliency Detection
- Crystallization and Solubility Studies
- X-ray Diffraction in Crystallography
- Human Pose and Action Recognition
- Advanced Computational Techniques and Applications
- Robotics and Sensor-Based Localization
- Air Quality and Health Impacts
- Image Processing and 3D Reconstruction
- Optical measurement and interference techniques
Marine Biomedical Research Institute of Qingdao
2025
Sichuan Academy of Medical Sciences & Sichuan Provincial People's Hospital
2018-2025
University of Electronic Science and Technology of China
2012-2025
Fujian Agriculture and Forestry University
2024-2025
Fujian Medical University
2025
Tencent (China)
2018-2024
Southwest University
2022-2024
Rutgers New Jersey Medical School
2024
KLA (United States)
2018-2024
Hubei University of Technology
2024
Versatile Video Coding (VVC) was finalized in July 2020 as the most recent international video coding standard. It developed by Joint Experts Team (JVET) of ITU-T Group (VCEG) and ISO/IEC Moving Picture (MPEG) to serve an ever-growing need for improved compression well support a wider variety today’s media content emerging applications. This paper provides overview novel technical features new applications core technologies achieving significant bit rate reductions neighborhood 50% over its...
Video anomaly detection under weak labels is formulated as a typical multiple-instance learning problem in previous works. In this paper, we provide new perspective, i.e., supervised task noisy labels. such viewpoint, long cleaning away label noise, can directly apply fully action classifiers to weakly detection, and take maximum advantage of these well-developed classifiers. For purpose, devise graph convolutional network correct Based upon feature similarity temporal consistency, our...
Image inpainting techniques have shown significant improvements by using deep neural networks recently. However, most of them may either fail to reconstruct reasonable structures or restore fine-grained textures. In order solve this problem, in paper, we propose a two-stage model which splits the task into two parts: structure reconstruction and texture generation. first stage, edge-preserved smooth images are employed train reconstructor completes missing inputs. second based on...
Generating portrait images by controlling the motions of existing faces is an important task great consequence to social media industries. For easy use and intuitive control, semantically meaningful fully disentangled parameters should be used as modifications. However, many techniques do not provide such fine-grained controls or indirect editing methods i.e. mimic other individuals. In this paper, a Portrait Image Neural Renderer (PIRenderer) proposed control face with three-dimensional...
In point cloud compression, sufficient contexts are significant for modeling the distribution. However, gathered by previous voxel-based methods decrease when handling sparse clouds. To address this problem, we propose a multiple-contexts deep learning framework called OctAttention employing octree structure, memory-efficient representation Our approach encodes symbol sequences in lossless way gathering information of sibling and ancestor nodes. Expressly, first represent clouds with to...
With the emerging applications such as online gaming and Wi-Fi display, screen content video, including computer generated text, graphics animations, becomes more popular than ever. Traditional video coding technologies typically were developed based on models that fit into natural, camera-captured video. The distinct characteristics exhibited between these two types of contents necessitate exploration efficiency improvement given new tools can be specially for HEVC Screen Content Coding...
An explainable machine learning method for point cloud classification, called the PointHop method, is proposed in this work. The consists of two stages: 1) local-to-global attribute building through iterative one-hop information exchange, and 2) classification ensembles. In stage, we address problem unordered data using a space partitioning procedure developing robust descriptor that characterizes relationship between its neighbor unit. When put multiple units cascade, attributes will grow...
Generic object detection algorithms have proven their excellent performance in recent years. However, on underwater datasets is still less explored. In contrast to generic datasets, images usually color shift and low contrast; sediment would cause blurring images. addition, creatures often appear closely each other due living habits. To address these issues, our work investigates augmentation policies simulate overlapping, occluded blurred objects, we construct a model capable of achieving...
This paper presents the intra prediction and mode coding of Versatile Video Coding (VVC) standard. standard was collaboratively developed by Joint Experts Team (JVET). It follows traditional architecture a hybrid block-based codec that also basis previous standards. Almost all features VVC either contain substantial modifications in comparison with its predecessor H.265/HEVC or were newly added. The key aspects these tools are following: 65 angular modes block shape-adaptive directions 4-tap...
Instance segmentation in high-resolution (HR) remote sensing imagery is one of the most challenging tasks and more difficult than object detection semantic tasks. It aims to predict class labels pixel-wise instance masks locate instances an image. However, there are rare methods currently suitable for HR images. Meanwhile, it implement due complex background In this article, a novel approach based on Cascade Mask R-CNN proposed, which called high-quality network (HQ-ISNet). scheme, HQ-ISNet...
In the past decade, development of transform coding techniques has achieved significant progress and several advanced tools have been adopted in new generation Versatile Video Coding (VVC) standard. this paper, a brief history during VVC standardization is presented, standard are described detail together with their initial design, incremental improvements implementation aspects. To improve efficiency, four introduced VVC, which namely Multiple Transform Selection (MTS), Low-Frequency...
<h3>Objective</h3> Precise genetic analyses were conducted with ring finger protein 213 (<i>RNF213</i>) in relation to a particular clinical phenotype Chinese patients moyamoya disease (MMD) determine whether heterozygosity is responsible for the early-onset and severe form of this disease. <h3>Methods</h3> A case–control study <i>RNF213</i> p.R4810K involving 1,385 MMD 2,903 normal control participants was performed. Correlation between genotype or different features also statistically...
The previous deep video compression approaches only use the single scale motion compensation strategy and rarely adopt mode prediction technique from traditional standards like H.264/H.265 for both residual compression. In this work, we first propose a coarse-to-fine (C2F) framework better compensation, in which perform estimation, twice coarse to fine manner. Our C2F can achieve results without significantly increasing bit costs. Observing hyperprior information (i.e., mean variance values)...
A steady increase in available processing power continues to drive advances video compression technology. The recently completed Versatile Video Coding (VVC) standard aims double the efficiency of HEVC and deliver a same quality at half bitrate. To achieve this goal, VVC includes several new methods that improve coding cost increased complexity. This paper provides complexity analysis its VTM reference software. Whereas is more complex than HEVC, it remains readily implementable software on...
Inspired by the recent PointHop classification method, an unsupervised 3D point cloud registration called R-PointHop, is proposed in this work. R-PointHop first determines a local reference frame (LRF) for every using its nearest neighbors and finds attributes. Next, obtains local-to-global hierarchical features downsampling, neighborhood expansion, attribute construction dimensionality reduction steps. Thus, correspondences are built feature space neighbor rule. Afterwards, subset of...
Mapping tree crown is critical for estimating the functional and spatial distribution of ecosystem services. However, accurate up-to-date urban mapping remains a challenge due to time-consuming nature field sampling heterogeneity. Another data cost, which always concern low-cost processing forest maps on large scales. Here, we developed novel working framework by integrating an advanced deep learning technology, Mask Region-based Convolutional Neural Network (Mask R-CNN) model with Google...
We deal with the controllable person image synthesis task which aims to re-render a human from reference explicit control over body pose and appearance. Observing that images are highly structured, we propose generate desired by extracting distributing semantic entities of images. To achieve this goal, neural texture extraction distribution operation based on double attention is described. This first extracts textures feature maps. Then, it distributes extracted according spatial...
With the rapid development of 3D vision, point cloud has become an increasingly popular visual media content. Due to irregular structure, posed novel challenges related research, such as compression, transmission, rendering and quality assessment. In these latest researches, assessment (PCQA) attracted wide attention due its significant role in guiding practical applications, especially many cases where reference is unavailable. However, current no-reference metrics which based on prevalent...
Block-based compression scheme shows remarkable success in image and video coding. However, existing tree-type block partition methods usually divide point clouds into clusters with few or disjoint points due to the irregular sampling, which is adverse for subsequent transform exploit local region correlation. Moreover, widely-used optimal transform, e.g., Discrete Cosine Transform (DCT), deduced assumption of specified probability model. Thus these transforms cannot adequately conform...
There is an urgent need from various multimedia applications to efficiently compress point clouds. The Moving Picture Experts Group has released a standard platform called geometry-based cloud compression (G-PCC). However, its <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">k</i> -nearest neighbor (k-NN) based attribute prediction limited efficiency for clouds with rich texture and directional information. To overcome this problem, we propose...
The substantial data volume within dynamic point clouds representing three-dimensional moving entities necessitates advancements in compression techniques. Motion estimation (ME) is crucial for reducing cloud temporal redundancy. Standard block-based ME schemes, which typically utilize the previously decoded as inter-reference frames, often yield inaccurate and translation-only estimates clouds. To overcome this limitation, we propose an advanced patch-based affine scheme geometry...
There is a pressing need across various applications for efficiently compressing point clouds. While the Moving Picture Experts Group introduced geometry-based cloud compression (G-PCC) standard, its attribute scheme falls short of eliminating signal frequency-domain redundancy. This paper proposes texture-guided graph transform optimization compression. We formulate coding task as problem, considering both decorrelation capability and sparsity optimized within tailored joint framework....