Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks359 from public data
- (Cross-)Browser Fingerprinting via OS and Hardware Level Features207
This paper proposes a browser fingerprinting technique that can track users not only within a single browser but also across different browsers on the same machine, and can achieve higher uniqueness rate than the only cross-browser approach in the literature with similar stability.
- Region-aware Adaptive Instance Normalization for Image Harmonization154
This paper proposes a simple yet effective Region-aware Adaptive Instance Normalization (RAIN) module, which explicitly formulates the visual style from the background and adaptively applies them to the foreground.
- 108
- IdentityDP: Differential private identification protection for face images105
IdentityDP, a face anonymization framework that combines a data-driven deep neural network with a differential privacy (DP) mechanism, is proposed that can effectively obfuscate the identity-related information of faces, preserve significant visual similarity, and generate high-quality images that can be used for identity-agnostic computer vision tasks, such as detection, tracking, etc.
- Freestyle Layout-to-Image Synthesis95
This work introduces a new module called Rectified Cross-Attention (RCA) that can be conveniently plugged in the diffusion model to integrate semantic masks, and opts to leverage large-scale pre-trained text-to-image diffusion models to achieve the generation of unseen semantics.
- Accuracy assessment of human trunk surface 3D reconstructions from an optical digitising system94
A commercial, optical 3D digitising system for the 3D reconstruction of the entire trunk for clinical assessment of external asymmetry is assessed and the resulting surface is a textured, high-density polygonal mesh.
- Personalized and Invertible Face De-identification by Disentangled Identity Information Manipulation87
- Semi-automatic computer aided lesion detection in dental X-rays using variational level set79
A semi-automatic lesion detection framework to detect areas of lesions from periapical dental X-rays using level set method that is able to segment the image into pathological meaningful regions for further analysis and automatically locate the PL and BL with a seriousness level marked for further dental diagnosis.
- High precision camera calibration in vision measurement72
In order to ensure the measurement precision for shape recovery from a stereo vision system, a novel high precision calibration method is presented.
- Contourlet-based image adaptive watermarking71
The proposed contourlet-based image adaptive water marking (CIAW) scheme is particularly superior to the conventional watermarking schemes when the watermarked image is attacked by some image processing methods, which destroy the HF subbands, thanks to theWatermarking components preserved in the LF subbands.
- Deep RNNs for video denoising70
A deep architecture based on Recurrent Neural Network (RNN) based on a patch-based end-to-end mapping between the clean and noisy video sequences that is able to compete with some existing video denoising methods.
- 67
- OmniVTLA: Vision-Tactile-Language-Action Models With Semantic-Aligned Tactile Sensing64
This paper proposes OmniVTLA, a novel framework that models vision, tactile, and language for end-to-end contact-rich manipulation tasks, and introduces ObjTac, a comprehensive tactile dataset comprising textual, visual, and tactile information for 56 objects across 10 categories.
- 63
- 59
- 56
- 360-Degree Panorama Generation from Few Unregistered NFoV Images54
A novel pipeline called PanoDiff is proposed, which efficiently generates complete 360° panoramas using one or more unregistered NFoV images captured from arbitrary angles and achieves state-of-the-art panoramic generation quality and high controllability, making it suitable for applications such as content editing.
- 54
- Boosting Video Object Segmentation via Space-Time Correspondence Learning51
This work devise a correspondence-aware training framework, which boosts matching-based VOS solutions by explicitly encouraging robust correspondence matching during network learning, and reinforces the standard, fully supervised training of mask segmentation with label-free, contrastive correspondence learning.
- 48
- 46
- 44
- Analysis and identification of spamming behaviors in Sina Weibo microblog44
This paper presents a systematic research on the analysis of spamming in Sina Weibo platform, which is currently a dominant microblogging service provider in China, and designs an automatic online spammer identification system.
- 41
- 40
- 4DGC: Rate-Aware 4D Gaussian Compression for Efficient Streamable Free-Viewpoint Video39
4DGC is proposed, a novel rate-aware 4D Gaussian compression framework that significantly reduces storage size while maintaining superior RD performance for FVV and introduces a motion-aware dynamic Gaussian representation that utilizes a compact motion grid combined with sparse compensated Gaussians to exploit inter-frame similarities.
- The Tenth NTIRE 2025 Efficient Super-Resolution Challenge Report39
This paper presents a comprehensive review of the NTIRE 2025 Challenge on Single-Image Efficient Super-Resolution, emphasizing groundbreaking advancements in state-of-the-art single-image ESR techniques.
- 39
- 38
- 38
- 37
- 37
- NTIRE 2021 Challenge on Video Super-Resolution36
This paper presents evaluation results from two competition tracks as well as the proposed solutions to the NTIRE 2021 Challenge on Video Super-Resolution, and develops conventional video SR methods focusing on the restoration quality.
- 33
- LongCat-Flash-Omni Technical Report31
This work introduces LongCat-Flash-Omni, a state-of-the-art open-source omni-modal model with 560 billion parameters, excelling at real-time audio-visual interaction and develops a modality-decoupled parallelism scheme specifically designed to manage the data and model heterogeneity inherent in large-scale multimodal training.
- 31
- 30
- 29
- 29
- 29
- 28
- 28
- Learning an Inverse Tone Mapping Network with a Generative Adversarial Regularizer28
A novel deep learning-based iTM method, which learns an inverse tone mapping network with a generative adversarial regularizer, and uses a U-Net-based HDR image generator to transfer input LDR images to HDR ones, and a simple CNN-based discriminator to classify the real HDR images and the generated ones.
- Diff-Restorer: Unleashing Visual Prompts for Diffusion-Based Universal Image Restoration27
Diff-Restorer is proposed, a diffusion-based universal image restoration framework that unifies diverse degradation handling within a single model and outperforms state-of-the-art methods in terms of generality, realism, and fidelity.
- Automatic Labanotation Generation Based on Human Motion Capture Data27
This article aims to generate Labanotation automatically from human motion capture data stored in BVH (Bio-vision Hierarchy) files by convertingMotion capture data into position format and building a sequence of coded description of Laban rotation.
- 26
- 26
- 23
- 23
- 23
- 23
- Background subtraction based on phase feature and distance transform23
A novel background subtraction method that can work under complex environments including dynamic background and illumination variations, especially for sudden illumination change and has no bootstrapping limitations.
- StableFace: Analyzing and Improving Motion Stability for Talking Face Generation22
Three effective solutions are proposed: a Gaussian-based adaptive smoothing module to smooth the 3D face representations to eliminate jitters in the input, and an objective metric (Motion Stability Index, MSI) to quantitatively measure the motion jitters.
- 22
- 21
- 21
- Implicit-Explicit Integrated Representations for Multi-View Video Compression20
This work proposes an implicit-explicit integrated representation for multi-view video compression that can achieve comparable or even superior performance to the latest multi-view video compression standard MIV and other INR-based schemes in terms of view compression and scene modeling.
- Generative Compression for Face Video: A Hybrid Scheme20
A bitrate-adjustable hybrid compression scheme for face video that combines the pixel-level precise recovery capability of traditional coding with the generation capability of deep learning based on abridged information, which proves the effectiveness of the proposed scheme.
- 19
- QWT: Retrospective and New Applications19
This paper provides a retrospective of QWT and investigates its potential use in the domain of image registration, image fusion, and color image recognition, and incorporates directional filter bank (DFB) into the quaternion wavelet decomposition scheme to greatly enhance the direction selectivity and anisotropy of Q WT.
- 19
- 18
- 18
- 18
- 18
- SSP-IR: Semantic and Structure Priors for Diffusion-Based Realistic Image Restoration17
This work integrates the visual comprehension capabilities of Multimodal Large Language Models and the visual representations of the original image to acquire accurate semantic prior, and introduces a Processor with RGB and FFT constraints to extract degradation-independent structure prior from the low-quality images.
- 17
- 17
- 17
- 16
- Deep Binary Representation for Efficient Image Retrieval16
Experiments on standard image retrieval benchmarks show that the proposed deep hashing method outperforms other state-of-the-art methods including unsupervised, supervised, and deep hashing methods.
- Biorthogonal multiple wavelets generated by vector refinement equation16
This paper provides a general method for the construction of compactly supported biorthogonal multiple wavelets by refinable function vectors which are the solutions of vector refinement equations of the form of a finitely supported sequence of r × r matrices called the refinement mask.
- Face De-Identification: State-of-the-Art Methods and Comparative Studies15
This survey provides a comprehensive review of state-of-the-art face de-identification methods, categorized into three levels: pixel-level, representation-level, and semantic-level techniques, demonstrating that deep learning-based approaches, particularly those using Generative Adversarial Networks and diffusion models, have achieved significant advancements in balancing privacy and utility.
- High-Fidelity Face Reenactment Via Identity-Matched Correspondence Learning15
A novel face reenactment framework that achieves both high-fidelity generation and identity preservation and outperforms state-of-the-art approaches both qualitatively and quantitatively is proposed.
- 15
- 15
- 15
- 14
- 14
- 14
- Efficient Dynamic-NeRF Based Volumetric Video Coding with Rate Distortion Optimization13
This work decomposes the NeRF representation into the coefficient fields and the basis fields, incrementally updating the basis fields in the temporal domain to achieve dynamic modeling and performs end-to-end joint optimization on the modeling and compression process to further improve the compression efficiency.
- 13
- 13
- 13
- 13
- 13
- 12
- Reorder user's tweets12
This work presents a supervised learning method for personalized tweets reordering based on user interests, and model the relevance of tweets by minimizing the pairwise loss of relevant and irrelevant tweets.
- 11
- 11
- FACT: Fused Attention for Clothing Transfer with Generative Adversarial Networks11
A novel semantic-based Fused Attention model for Clothing Transfer (FACT), which allows fine-grained synthesis, high global consistency and plausible hallucination in images, and develops a stylized channel-wise attention module to capture correlations on feature levels.
- 11
- 11
- 11
- 11
- Subgraphs Matching-Based Side Information Generation for Distributed Multiview Video Coding11
In order to find a distinctive feature matching with a more stable approximation, linear and nonlinear projections are adopted to reduce the dimension SIFT descriptors, and TPS (thin plate spline) warping model is to catch a more accurate interview motion model.
- Contourlet Image Coding Based on Adjusted SPIHT11
This paper proposes a contourlet image coding algorithm by constructing a virtual low frequency subband and adjusting coding method of SPIHT (Set Partitioning in Hierarchical Trees) algorithm according to the structure ofcontourlet coefficients.
- Rate-aware Compression for NeRF-based Volumetric Video10
This paper tries to directly learn a compact NeRF representation for volumetric video in the training stage based on the proposed rate-aware compression framework, and proposes an adaptive quantization strategy and learns the optimal quantization step for the NeRF representations.
- 10
- 10
- 10
- Audio Turing Test: Benchmarking the Human-likeness of Large Language Model-based Text-to-Speech Systems in Chinese8
The Audio Turing Test (ATT), a multi-dimensional Chinese corpus dataset ATT-Corpus paired with a simple, Turing-Test-inspired evaluation protocol, which asks evaluators to judge whether a voice sounds human, reduces rating bias and improves evaluation robustness.
- AnimeColor: Reference-based Animation Colorization with Diffusion Transformers8
This work proposes AnimeColor, a novel reference-based animation colorization framework leveraging Diffusion Transformers (DiT), which integrates sketch sequences into a DiT-based video diffusion model, enabling sketch-controlled animation generation.
- Pioneer: Offline Reinforcement Learning based Bandwidth Estimation for Real-Time Communication8
Pioneer eliminates the need for real-time environment interaction during training and achieves good performance through lightweight training, and test results on offline datasets show that Pioneer can achieve better performance than expert algorithms.
- Achieving Privacy-Preserving Multi-View Consistency with Advanced 3D-Aware Face De-identification8
This work focuses on identity disentanglement within the latest 3D-aware face generation model, and proposes an advanced face de-identification framework that can be applied to various scenarios and results in high-quality, identity-consistent de-identification that preserves other facial features.
- 8
- Learning Dense UV Completion for Human Mesh Recovery8
This paper proposes Dense Inpainting Human Mesh Recovery (DIMR), a two-stage method that leverages dense correspondence maps to handle occlusion and designs a feature inpainting training procedure that guides the network to learn from unoccluded features.
- 8
- 8
- 8
- 8
- Fast and Robust Clinical Triple-Region Image Segmentation Using One Level Set Function8
Results show that the proposed method is able to successfully segment the triple-region using a single level set function, and this segmentation is very robust to the placement of initial contour.
- 7
- 7
- 7
- Intra Encoding Complexity Control with a Time-Cost Model for Versatile Video Coding7
Experimental results show that, for the challenging intra coding scenario, the complexity error quickly converges to under 3.21%, while keeping a reasonable time saving and rate-distortion (RD) performance, which proves the efficiency of the proposed methods.
- 7
- 7
- 7
- 7
- 7
- A level set segmentation for computer-aided dental x-ray analysis7
Experimental results show that the proposed framework achieves faster level set segmentation and provides more detailed information and indications of possible problems to the dentist.
- 6
- 6
- 6
- 6
- 6
- 6
- 6
- 6
- 6
- 6
- BEST: Benchmark and Evaluation of Surveillance Task6
Compared with the existing surveillance/non-surveillance datasets, the proposed BEST dataset provides a realistic, extensive and diversified testbed for a more comprehensive performance evaluation, and experimental results show that, performance of seven pedestrian detection algorithms on BEST is worse than that on the existing datasets.
- 6
- 6
- Multi-illumination Face Recognition from a Single Training Image per Person with Sparse Representation6
This paper extends Sparse Representation Classification (SRC) to the one sample per person problem by introducing relighting methods to generate virtual faces and verifying the recognition under different lighting environments by a cross-database comparison.
- 6
- 6
- 6
- 6
- Eye Location Using Gabor Transform6
Experimental results show that the proposed method to locate eyes using Gabor transform is fairly robust to variations of expression, pose and illumination.
- 6
- 5
- 5
- 5
- Generative Architecture for Data Imputation in Secure Blockchain-enabled Spatiotemporal Data Management5
A groundbreaking blockchain-enabled evolutionary indirect feedback graph algorithm for the secure management of interconnected spatiotemporal datasets, utilizes a generative neural network model for data imputation and incorporates indirect feedback mechanisms, allowing users to provide implicit feedback through their interactions, enhancing the relevance and efficiency of data retrieval.
- Local Bidirection Recurrent Network for Efficient Video Deblurring with the Fused Temporal Merge Module5
This work adopts the recurrent-based method as the baseline and reconsider the recurrent mechanism as well as the temporal feature alignment in the state-of-the-art methods, and adds the local backward connection to the global forward recurrent backbone to effectively exploit accurate future information.
- 5
- 5
- 5
- 5
- Light field image restoration in low-light environment5
A high-dimensional convolutional neural network is introduced to extract the spatio-angular features from the LF to reduce color distortion and achieve noticeable improvement compared with state-of-the-art low-light image restoration techniques in both visual inspection and objective assessments.
- 5
- 5
- 5
- 5
- 5
- 5
- 5
- 4
- 4
- Serial Low-rank Adaptation of Vision Transformer4
Serial LoRA is proposed, a novel LoRA variant that introduces a shared low-rank matrix serially composite with the attention mechanism that extracts the underlying commonality of parameters in adaptation, significantly reducing redundancy.
- Enhanced Semantic Extraction and Guidance for UGC Image Super Resolution4
This paper proposes a novel approach to UGC image super-resolution by integrating semantic guidance into a diffusion framework and enhances degradation removal and detail generation by incorporating a pretrained semantic extraction model (SAM2) and fine-tuning key hyperparameters for improved perceptual fidelity.
- PA-HOI: A Physics-Aware Human and Object Interaction Dataset4
This work introduces the PA-HOI Motion Capture dataset, which highlights the impact of objects' physical attributes on human motion dynamics, including human posture, moving velocity, and other motion characteristics, and extends the scope of existing data sets for understanding how the physical attributes of different objects influence human posture, speed, motion scale, and interacting strategies.
- 4
- Multi-Scale Coarse-to-Fine Transformer for Frame Interpolation4
This work proposes a two-stage flow-free video interpolation architecture that outperforms state-of-the-art methods for both single frame and multi frames interpolation tasks, and the extended ablation studies verify the effectiveness of the model.
- Multi-task Semi-supervised Learning Framework for Surgical Instrument Pose Estimation4
A semi-supervised and multi-task learning based pose estimation framework to enhance the model effect using unannotated data and overcome the problem of insufficient data in surgical instrument pose estimation task is proposed.
- 4
- Real-Time End-to-End Monaural Multi-Speaker Speech Recognition4
This paper comprehensively investigate and compare the mainstream end-to-end multi-speaker speech recognition systems, and improves the recently proposed non-autoregressive end-to-end speech recognition model Mask-CTC, and introduces it to multi-speaker speech recognition to achieve real-time decoding.
- Multi-scale Generative Adversarial Learning for Facial Attribute Transfer4
The Structural Similarity Index (SSIM) is introduced in GAN training objective as the measurement between input images and reconstructed images to facilitate feature learning and encourage long-term correlation.
- 4
- Real-time UHD video super-resolution and transcoding on heterogeneous hardware4
This paper optimize the most time consuming task, video SR, with CUDNN and CUDA libraries to achieve real-time processing speed for a single UHD output frame on an ordinary GPU, and designs a GPU-friendly multi-thread scheduling algorithm for data and computation to better utilize GPU resources and achieve real -time performance on outputting UHD video clips.
- 4
- 4
- 4
- 4
- Video Coding With Key Frames Guided Super-Resolution4
In this paper a video coding scheme with Layered Block Matching Super-resolution (LBM-SR) is presented, which utilizes a Layered block Matching method in wavelet domain to restore the lost high frequency parts of the non-key frame with the nearby key frames as a reference.
- 4
- 4
- 4
- 4
- Auto-detection of strip area in 3D measurement4
The aim of this research is to find a method to auto-detect the strip area, which can save a lot of time for the 3D measurement, automatically complete the3D recovery procedure, and bring convenience to operators.
- FreeInsert : Personalized Object Insertion with Geometric and Style Control3
FreeInsert, a novel training-free framework that customizes object insertion into arbitrary scenes by leveraging 3D geometric information, is proposed, producing geometrically controlled, style-consistent edited images via the diffusion model.
- Consistent Video Colorization via Palette Guidance3
A palette-based color guider is designed to assist the model in generating vivid and consistent colors and demonstrates that the proposed method can provide vivid and stable colors for videos, surpassing previous methods.
- Multimodal Semantic-Aware Automatic Colorization with Diffusion Prior3
Experiments indicate that the proposed automatic colorization pipeline synthesizes saturated colors while maintaining plausible semantics, surpassing previous methods in terms of perceptual realism and gain most human preference.
- SAFR: A Real-Time Communication System with Adaptive Frame Rate3
This paper proposes a novel RTC system, SAFR (System of Adaptive Frame Rate), to jointly optimize all the key indicators of QoE performance, and shows greater superiority under worse network conditions.
- 3
- 3
- 3
- A Deep Tracking and Segmentation Approach for Soccer Videos Visual Effects3
A novel deformable cross-similarity correlation (DF_CORR) is adopted to estimate the deformation of players and the model outperforms the state-of-the-art trackers whose accuracy is decreased significantly compared with the general tracking tasks.
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- Neural network prediction model for ship hydraulic pressure signal under wind wave background3
An effectual way to detect the ship using the ship hydraulic pressure signal is presented and the practical result shows that the method can improve the signal to noise ratio (SNR) obviously.
- Texture Direction Based Optimization for Intra Prediction in HEVC3
Experimental results show that the proposed algorithm provides an average 40% total encoding time reduction with the negligible loss of rate-distortion.
- 3
- 3
- 3
- MCM: An Efficient Geometric Constraint Method for Robust Local Feature Matching3
Results show that the new geometric constraint method Maximal Clique Matching (MCM) is robust to occlusion, clutter, deformations and repetitive patterns, and outperforms other up-to-date techniques.
- 3
- 3
- 3
- Rate-Aware Learned Speech Compression2
A rate-aware learned speech compression scheme that replaces the quantizer with an advanced channel-wise entropy model to improve RD performance, simplify training, and avoid codebook collapse is proposed.
- Semantic and Temporal Integration in Latent Diffusion Space for High-Fidelity Video Super-Resolution2
Semantic and Temporal Guided Video Super-Resolution (SeTe-VSR) is proposed, a novel approach that incorporates both semantic and temporal-spatio guidance in the latent diffusion space to address challenges of complex video super-resolution tasks.
- 2
- Approaches and Challenges for Resolving Different Representations of Fictional Characters for Chinese Novels2
A state-of-the-art anaphora resolution model is adapted to resolve character representations in Chinese novels by making some modifications, and a widely used BERT fine-tuned model for speaker extraction as assistance is trained.
- 2
- An improved resource allocation method for mapping service function chains based on A3C2
An improved service function chain mapping resource allocation method (SA3C) based on the Asynchronous advantageous action evaluation algorithm (A3C) is proposed and a mathematical model for joint allocation, which modeled the minimization of processing time as a Markov process is proposed.
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- Current Frame Priors Assisted Neural Network for Intra Prediction2
A Current Frame Priors assisted Neural Network (CFPNN) is presented to improve the intra prediction efficiency and the best way to integrate the proposed method into the codec is investigated.
- 2
- 2
- 2
- Improving Semantic Style Transfer Using Guided Gram Matrices2
This work investigates semantic style transfer for content images with more than 2 semantic regions by combining guided Gram matrices with gradient capping and multi-scale representations, which simplifies the parameter tuning problem, improves the style transfer results and is faster than current semantic methods.
- Missing pins detection for power equipment firmware using unmanned aerial vehicle images2
Experiments show that the proposed algorithm of missing pins detection based on UAV image can achieve competitive results to detect the defects in special scenes, and has good robustness, which satisfies the real-time and accuracy requirements.
- 2
- Spatial and Temporal Feature-Based Reduced Reference Quality Assessment for Rate-Varying Videos in Wireless Networks2
A spatial and temporal feature-based reduced reference (RR) quality assessment for rate-varying videos in wireless networks called STRQAW is proposed, which reflects human subjective feelings well and it provides an evaluation for adjusting compression-coding rates in real time.
- 2
- 2
- A Medium Access Control Mechanism for Distributed In-band Full-Duplex Wireless Networks2
A medium access control mechanism for use in distributed IBFD wireless networks, FD-DMAC (Full-Duplex Distributed MAC), which can assist communication nodes to adjust transmit powers and suppress INI and is given a theoretical analysis of network performance using a discrete-time Markov model.
- 2
- 2
- 2
- Research and implementation of simulation for TDICCD remote sensing in vibration of optical axis2
It is quite significant for improving simulation accuracy and efficiency to give an appropriate discrete modeling and simulation method, which is based on the principle of TDICCD sensor, and the gray value of pixels in sensor matrix is obtained by a weighted arithmetic, which solves the problem of pixels dismatch.
- 2
- A Novel Digital Watermarking Algorithm Based on Wavelet Lifting Scheme and Linear Regression2
The simulation results show that watermarked image can obtain a large value of PSNR and demonstrate that the algorithm has strong robustness to resist noise and clipping attacks, it also can greatly reduce carrier image distortion.
- Piecewise FARIMA models for long-memory time series2
A four-step procedure is proposed to find out the BPs and to estimate the parameters of each regime and its effectiveness is shown by Monte Carlo simulations and an application to real traffic data modelling is considered.
- Improving Detector of Viola and Jones through SVM2
A new cascading structure added SVM stages which employ the confidence values of multiple preceding Adaboost stages as input is presented, and a rejection hyperplane and a promotion hyperplane are learned for eachadded SVM stage.
- 2
- 3D Digitizing Device Applied in Evaluation and Simulation of Postoperative Trunk Surface Shape in Adolescent Idiopathic Scoliosis2
A hybrid, deformable model of the human trunk is presented for prediction of surgical outcome on trunk shape in AIS and a validated simulation tool is developed to develop a validated simulation tool that would allow the clinician to illustrate to the patient the potential result of the surgery and help in deciding on a surgical strategy that could most improve their external appearance.
- 2
- 2
- Adaptive Predict Based on Fading Compensation for Lifting-Baed Motion Compensated Temporal Filtering2
- A Hybrid Scheme for Face Video Compression1
This article proposes a hybrid compression scheme for face videos which combines traditional coding with generative compression, and devise a deep video generation model to synthesize smooth video frames according to the extracted sparse keypoints.
- AI-HEAD: Auto-indexing for Helical assembles’ diffraction and fast dynamic structure determination1
Automatic, fast and accurate helical parameter and structure calculation method AI-HEAD, which perform diffraction indexing and parameter calculation automatically in seconds, and determine the final structure unambiguously by sorting parameters semi-automatically is developed.
- MultiEgo: A Multi-View Egocentric Video Dataset for 4D Scene Reconstruction1
Experiment validation demonstrates the practical utility and effectiveness of the MultiEgo dataset for free-viewpoint video (FVV) applications, establishing MultiEgo as a foundational resource for advancing multi-view egocentric dynamic scene reconstruction research.
- 1
- MoRLACS: A Monocular RGBD-based Locomotion Approach for CAVE Systems1
MoRLACS, a novel locomotion approach for CAVE systems that leverages a single RGBD camera, integrates small-scale physical walking with controller-based large-scale exploration through a tailored guidance method and supports extended interaction in larger virtual environments.
- 1
- TacCompress: A Benchmark for Multi-Point Tactile Data Compression in Dexterous Hand1
This paper introduces a Multi-Point Tactile Dataset for Dexterous Hand Grasping (Dex-MPTD), and investigates both lossless and lossy compression on Dex-MPTD by converting tactile data into images and applying six lossless and five lossy image codecs for efficient compression.
- A New People-Object Interaction Dataset and NVS Benchmarks1
A new people-object interaction dataset that comprises 38 series of 30-view multi-person or single- person RGB-D video sequences, accompanied by camera parameters, foreground masks, SMPL models, some point clouds, and mesh files is introduced.
- 1
- 1
- 1
- 1
- ViCoFace: Learning Disentangled Latent Motion Representations for Visual-Consistent Face Reenactment1
This work proposes an effective and cost-efficient face reenactment approach that disentangles the full-head motion into two parts: the transferable motion and preservable motion and then compose the full motion representation using latent motions from the source image and the target image.
- 1
- 1
- NeRF-SDP: Efficient Generalizable Neural Radiance Field with Scene Depth Perception1
This work proposes a novel framework, NeRF-SDP, which achieves both efficiency and high fidelity by introducing scene depth perception, and achieves comparable synthesis quality to state-of-the-art methods while significantly improving rendering efficiency.
- 1
- 1
- 1
- 1
- 1
- Blindly Predict Image and Video Quality in the Wild1
This work proposes a two-stage architecture to separately predict image and video quality in the wild, and develops a Relation-Guided Temporal Attention module for video quality prediction, which captures global inter-frame dependencies in embedding space to learn frame-wise attention weights for frame quality aggregation.
- 1
- 1
- 1
- Reconstruction of Multi-view Video Based on GAN1
The VGG network is used to extract the high-dimensional features between the images, indicating the relevance of the adjacent images, and the GAN is further used to more accurately generate virtual viewpoint images.
- Long-Term Tracking Based on Multi-Feature Adaptive Fusion for Video Target1
A long-term tracking algorithm based on multi-feature adaptive fusion for video target is presented, which design a robust appearance model by fusing powerful features including histogram of gradient, local binary pattern and color-naming at response map level to conquer the interference in the video.
- 1
- 1
- 1
- 1
- Segmentation and Description of Human Facial Features Region1
MPEG-7 is an important criterion for the multimedia content descriptions, so it has a great application value for an in-depth study, according to the study of MPEG-7 standard construction of face related structure models.
- A Path Recognition Algorithm Based on Image Process for Smart Car1
Experiments show that the tracking movement of the smart car is real-time, efficient and intelligent, which has achieved the expected goals.
- 1
- 1
- A procedure for modeling non-stationary signals with long range dependence1
The problem of modeling non-stationary signals with long range dependence is considered by using piecewise fractional autoregressive integrated moving average processes and a procedure to find out all the parameters of the model is proposed.
- 1
- Spatiotemporal Phase Congruency Based Invariant Features for Human Behavior Classification1
The proposed model detects feature points by measuring the spatiotemporal phase congruency of Fourier series components along with their characteristic scale and principal orientation and has key advantages of this interest point detector, including the invariance to contrast variations and more precise feature location.
- Structure-Preserving Colorization Based on Quaternionic Phase Reconstruction1
A novel semiautomatic colorization method is proposed based on quaternionic phase reconstruction, whose polar magnitude and polar phase are recovered from the intensity of original grayscale image and color scribbles of user's manual input, so that effective color propagation from less manual input can be expected.
- 1
- –
- Generative Communications: Overview, Technologies, and Trends–
Generative communications (GenCom), a novel paradigm for 6G networks in which large AI models drive semantic understanding, reasoning, and content generation, embedding these into the communication process, is introduced.
- A fast video coding scheme based on perceptual rate distortion optimized preprocessing–
A perceptual quality oriented R-D optimization video coding framework based on interaction between pre-encoding processing and in-loop R-D optimized mode decision, and a mask-guided early skip decision mechanism that uses the masks generated by preprocessing to terminate further block partitioning during encoding process, bypassing redundant R-D optimized mode decision.
- –
- –
- Adaptive Bitrate Live Streaming over HTTP-FLV: A Practical System Perspective–
This paper presents the implementation and deployment experience of adaptive bitrate live streaming over HTTP-FLV in TikTok Live, and implements both client-executed and server-executed ABR switching, and compares them through large-scale online A/B tests.
- –
- –
- –
- –
- Stable recovery guarantees for blind deconvolution under random mask assumption–
This study addresses the blind deconvolution problem with modulated inputs, focusing on a measurement model where an unknown blurring kernel $\boldsymbol{h}$ is convolved with multiple random modulations of a signal $\boldsymbol{x}$, subject to $\ell_2$-bounded noise".
- –
- –
- –
- –
- –
- In-Context Translation: Towards Unifying Image Recognition, Processing, and Generation–
In-Context Translation (ICT), a general learning framework to unify visual recognition, low-level image processing, and conditional image generation, and edge-to-image synthesis, is proposed.
- –
- –
- –
- –
- –
- –
- Perceptual Video Coding Based on Spatial Masking for Medical Video Communication–
An adaptive quantization parameter algorithm based on fast spatial masking estimation to reduce bitrate while maintain near lossless perceptual quality and meets real-time requirements is proposed.
- –
- Improved features using convolution-augmented transformers for keyword spotting–
This paper seeks to combine convolution and Transformers improves over using them individually, and proposes improved features using convolution-augmented transformers for keyword spotting, showing that the improved features can yield at least 3% improvement compared with other features.
- 19.2: Accurate and Fast Eye‐box Measurement for Near‐eye Display Based on Image Recognition Algorithm–
A fast and accurate measurement scheme based on image recognition algorithm by an automatic robot measuring system is introduced for eye‐box measurement for near‐eye display.
- –
- –
- –
- –
- –
- –
- –
- –
- Multi-level Ponte osteotomies combined with four-rods compression in correcting Scheuermann's kyphosis–
Multi-level Ponte osteotomies combined with four-rods compression in correcting Scheuermann's kyphosis (SK) via a posterior approach is effective in dispersing stress of each rod and in achieving satisfied clinical outcomes, as well as maintaining correction of kYphosis and preventing implant-related complications.
- –
- –
- –
- –
- –
- Robust watermark algorithm for adaptive choice of information embedding position–
The improved robust watermark algorithm based on Support Vector Regression and Genetic Algorithm can resist watermarking attacks more effectively and it has high transparency under the situation with big capacity information embedded.
- –
- –
- Measurement Algorithm for Image Structure Noise on Hardcopy–
An algorithm for measuring image structure noise, including moire and screen dot effect, is described, showing the high correlation between human rating and metrics proves the effectiveness of the model.
- –
- Design of Testing System for Main Spindle Box of CNC Lathe Based on LabVIEW–
Testing system of hardware and software structure adopts sensors, data acquisition cards, industrial control computer, frequency converter and so on as hardware, good data analysis methods, and LabVIEW as software tool to realize real-time monitoring of main spindle box testing conditions.
- –
- Skew Estimation Based on Haar-Like Features–
Experimental results show that the skew estimation algorithm performs well on general printed documents with different contents, languages and layouts and the accuracy of skew angle estimation is comparable with or better than state-of-the-art methods.
- –
- –
- –
- Research on the 3D Error Analysis and Deformation Prediction for Part–
3D error analysis and deformation prediction system is proposed in this article, it has higher reliability and a certain role for the correct evaluation of processing conditions and comprehensive understanding of the process of evaluating the whole production process.
- –
- –
- –
- Personalized Reconstruction of 3D Face Based on Different Race–
The 3D face reconstruction method of different race chose different standard face model according to different race, and adjusted and combined with the extracted character and corresponding model, thereby acquired the personalized model that reflected different race.
- Image denoising using local tangent space alignment–
This work proposes a novel image denoising approach, which is based on exploring an underlying (nonlinear) lowdimensional manifold using local tangent space alignment (LTSA), and 'learn' such a manifold, which approximates the image content effectively.
- –
- –
- –
- LTE E2E Solution of Alcatel-Lucent Shanghai Bell–
Alcatel-Lucent Shanghai Bell's eNodeB uses common platform and supports 2G/3G smoothly evolution and compatible, evolved packet core need traffic management capability to ensure high QoS performance.
- Research on Floristics of Seed Plants in Hefu Forest Park–
This area is a transitional region from subtropical zone to warm temperate zone, and the composition,areal-types and floristics characteristics of seed plants were studied.
- –
- Method on real-time detection of ship static electric signature based on MESE–
A method that detecting the object based on maximum entropy spectral estimation (MESE) and according to the floating threshold, the object is detected by the moving power spectrum algorithm and practical results show that the method is effective.
- –
- –
- –
- –
- Tactical Skill in Prepositional Phrase of English-Chinese Translation–
The author sums up six rules and skills of understanding and translating English prepositional phrase, which is supposed to be helpful to ET and EL.
- Progressive refinement for robust image registration–
A new image registration algorithm with robust cost function and progressive refinement estimation is developed on the basis of direct method (DM) to avert larger local noise and outliers and get global optimization and reduce search space.
- –
- Research on expressing formulae in network systems–
FMLViewer was presented, which is a powerful tool for explaining and displaying FML and its usages has some value in both theory and practice for XML studies.
- –
- Formations and interpretations of some new English words and phrases–
The formations of the new words disscussed here are as follows:abbreviation, a new word is formed by abbreviating the initials of some words or phrases;combination, anew word is constituted by combining some old words or part of the old words; meaning addition,a new word comes into being by keeping the form of an old word but with a novel and different meaning.
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-12. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.