Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks381 from public data
- Relation-Aware Global Attention for Person Re-Identification654
This work proposes an effective Relation-Aware Global Attention (RGA) module which captures the global structural information for better attention learning and proposes to stack the relations, i.e., its pairwise correlations/affinities with all the feature positions together to learn the attention with a shallow convolutional model.
- Style Normalization and Restitution for Generalizable Person Re-Identification440
The aim of this paper is to design a generalizable person ReID framework which trains a model on source domains yet is able to generalize/perform well on target domains, and to enforce a dual causal loss constraint in SNR to encourage the separation of identity-relevant features and identity-irrelevant features.
- Image-to-Image Translation: Methods and Applications368
The key techniques of the existing I2I works are analyzed and clarified to clarify the main progress the community has made and point out remaining challenges in related fields.
- Inpaint Anything: Segment Anything Meets Image Inpainting334
The core idea behind IA is to combine the strengths of different models in order to build a very powerful and user-friendly pipeline for solving inpainting-related problems.
- Densely Semantically Aligned Person Re-Identification320
This work is the first to make use of fine grained semantics for addressing misalignment problems for re-ID and construct a set of densely semantically aligned part images (DSAP-images), where the same spatial positions have the same semantics across different person images.
- 257
- Diffusion Models for Image Restoration and Enhancement: A Comprehensive Survey224
This first to present a comprehensive review of recent diffusion model-based methods on image restoration, encompassing the learning paradigm, conditional strategy, framework design, modeling strategy, and evaluation, and proposes nine potential and challenging directions for the future research of diffusion model-based IR.
- Region Normalization for Image Inpainting207
It is shown that the mean and variance shifts caused by full-spatial FN limit the image inpainting network training and a spatial region-wise normalization named Region Normalization (RN) is proposed to overcome the limitation.
- Cloth-Changing Person Re-identification from A Single Image with Gait Prediction and Regularization200
This paper introduces Gait recognition as an auxiliary task to drive the Image ReID model to learn cloth-agnostic representations by leveraging personal unique and cloth-independent gait information, and introduces this framework as GI-ReID, which enables an efficient and latency-free person identity matching for surveillance.
- Fast integer-pel and fractional-pel motion estimation for H.264/AVC195
A hybrid Unsymmetrical-cross Multi-hexagon-grid Search (UMHexagonS) algorithm is introduced, which well solves the false motion vector estimation problem because of the local-minimum and can save 30–50% computation compared with the Full Fractional-pel Search scheme.
- Causal Contextual Prediction for Learned Image Compression190
A causal context model is proposed that separates the latents across channels and makes use of channel-wise relationships to generate highly informative adjacent contexts and a causal global prediction model to find global reference points for accurate predictions of undecoded points.
- GraphIQA: Learning Distortion Graph Representations for Blind Image Quality Assessment185
A Distortion Graph Representation (DGR) learning framework for IQA is proposed, in which each distortion is represented as a graph, i.e., DGR, and one can distinguish distortion types by learning the contrast relationship between these different DGRs, and can infer the ranking distribution of samples from different levels in a DGR.
- Task Residual for Tuning Vision-Language Models182
A new efficient tuning approach for VLMs named Task Residual Tuning (TaskRes), which performs directly on the text-based classifier and explicitly decouples the prior knowledge of the pre-trained models and new knowledge regarding a target task.
- Semantics-Aligned Representation Learning for Person Re-Identification164
A framework that drives the reID network to learn semantics-aligned feature representation through delicate supervision designs is proposed and achieves the state-of-the-art performances on the benchmark datasets CUHK03, Market1501, MSMT17, and the partial person reID dataset Partial REID.
- Blind Omnidirectional Image Quality Assessment With Viewport Oriented Graph Convolutional Networks162
This paper proposes a novel Viewport oriented Graph Convolution Network (VGCN) for blind omnidirectional image quality assessment (IQA) and demonstrates that the proposed model outperforms state-of-the-art full-reference and no-reference IQA metrics on two public omniddirectional IQA databases.
- Conditional Image-to-Image Translation147
This paper twists two conditional translation models together for inputs combination and reconstruction while preserving domain independent features and carries out experiments on men's faces from-to women's faces translation and edges to shoes&bags translations.
- Locating induced earthquakes with a network of seismic stations in Oklahoma via a deep learning method146
A fully convolutional network is developed and located earthquakes induced during oil and gas operations in Oklahoma with data from 30 network stations and the automated system is able to distinguish certain interfered events or events out of the monitoring zone based on the output probability estimate.
- 146
- Learning for Video Compression144
The proposed PixelMotionCNN (PMCNN) which includes motion extension and hybrid prediction networks can model spatiotemporal coherence to effectively perform predictive coding inside the learning network and provides a possible new direction to further improve compression efficiency and functionalities of future video coding.
- GraphAdapter: Tuning Vision-Language Models With Dual Knowledge Graph138
An effective adapter-style tuning strategy, dubbed GraphAdapter, which performs the textual adapter by explicitly modeling the dual-modality structure knowledge with a dual knowledge graph, which significantly outperforms previous adapter-based methods.
- Training-free Camera Control for Video Generation136
The capability of CamTrol to generalize to various base models, as well as its impressive applications in scalable motion control, dealing with complicated trajectories and unsupervised 3D video generation, are shown.
- MetaAlign: Coordinating Domain Alignment and Classification for Unsupervised Domain Adaptation135
This paper proposes an effective meta-optimization based strategy dubbed MetaAlign, which treats the domain alignment objective and the classification objective as the meta-train and meta-test tasks in a meta-learning scheme, and encourages both tasks to be optimized in a coordinated way.
- Towards Accurate One-Stage Object Detection With AP-Loss135
A novel optimization algorithm is developed, which seamlessly combines the error-driven update scheme in perceptron learning and backpropagation algorithm in deep networks, and good convergence property of the proposed algorithm is verified theoretically and empirically.
- Multi-Granularity Reference-Aided Attentive Feature Aggregation for Video-Based Person Re-Identification132
This paper proposes an attentive feature aggregation module, namely Multi-Granularity Reference-aided Attentive Feature Aggregation (MG-RAFA), to delicately aggregate spatio-temporal features into a discriminative video-level feature representation.
- Uncertainty-Aware Multi-Shot Knowledge Distillation for Image-Based Object Re-Identification127
This paper designs an Uncertainty-aware Multi-shot Teacher-Student (UMTS) Network that consists of a teacher network that learns the comprehensive features from multiple images of the same object, and a student network that takes a single image as input.
- Layer-Wise Coordination between Encoder and Decoder for Neural Machine Translation127
The concept of layer-wise coordination for NMT is proposed, which explicitly coordinates the learning of hidden representations of the encoder and decoder together layer by layer, gradually from low level to high level.
- 124
- Tensor Oriented No-Reference Light Field Image Quality Assessment122
The proposed Tensor-NLFQ model outperforms state-of-the-art 2D, 3D, multi-view, and LFI quality assessment algorithms and measures spatial-dimensional quality of LFI considering its global naturalness and local frequency properties.
- 116
- Style Normalization and Restitution for Domain Generalization and Adaptation113
A novel Style Normalization and Restitution module (SNR) is designed to simultaneously ensure high generalization and discrimination capability of the networks and enforce a dual restitution loss constraint to encourage the better separation of task-relevant and task-irrelevant features.
- No-Reference Light Field Image Quality Assessment Based on Spatial-Angular Measurement104
A No-Reference Light Field image Quality Assessment (NR-LFQA) scheme, where the main idea is to quantify the LFI quality degradation through evaluating the spatial quality and angular consistency, and the weighted local binary pattern is proposed to capture the characteristics of local angular consistency degradation.
- No-Reference Quality Assessment for 360-Degree Images by Analysis of Multifrequency Information and Local-Global Naturalness99
This model is the first no-reference quality assessment method for 360-degree images that combines multifrequency information and image naturalness and outperforms state-of-the-art full-reference (FR) and NR approaches.
- Learning to Run Challenge Solutions: Adapting Reinforcement Learning Methods for Neuromusculoskeletal Environments98
This work presents eight solutions that used deep reinforcement learning approaches, based on algorithms such as Deep Deterministic Policy Gradient, Proximal Policy Optimization, and Trust Region Policyoptimization, to make it run as fast as possible through an obstacle course.
- CCEdit: Creative and Controllable Video Editing via Diffusion Models97
This work presents CCEdit, a versatile generative video editing framework based on diffusion models that employs a novel trident network structure that separates structure and appearance control, ensuring precise and creative editing capabilities.
- Deep Frequency Filtering for Domain Generalization97
The effectiveness of the proposed Deep Frequency Filtering is demonstrated and it is shown that applying the DFF on a plain baseline out-performs the state-of-the-art methods on different domain generalization tasks, including close-set classification and open-set retrieval.
- 94
- 89
- LIQA: Lifelong Blind Image Quality Assessment87
This work proposes a novel Lifelong blind Image Quality Assessment (LIQA) approach, targeting to achieve the lifelong learning of BIQA, and adopts the Split-and-Merge distillation strategy to train a single-head network that makes task-agnostic predictions.
- Blind quality assessment for image superresolution using deep two-stream convolutional networks87
Experimental results on three publicly available SR image quality databases demonstrate the effectiveness and generalization ability of the proposed DeepSRQ method compared with state-of-the-art image quality assessment algorithms.
- Unsupervised Single Image Deraining with Self-Supervised Constraints87
An Unsupervised Deraining Generative Adversarial Network (UD-GAN) is proposed to tackle single image deraining problems by introducing self-supervised constraints from the intrinsic statistics of unpaired rainy and clean images.
- 80
- Dense Interaction Learning for Video-based Person Re-identification77
A hybrid framework that takes the principal advantages of both CNN-based and Attention-based architectures to tackle video-based person re-ID difficulties and introduces Spatio-TEmporal Positional Embedding into the DI decoder to investigate the positional relation among the spatial-temporal inputs.
- Global Distance-Distributions Separation for Unsupervised Person Re-identification75
A global distance-distributions separation (GDS) constraint over the two distributions is introduced to encourage the clear separation of positive and negative samples from a global view and validation of the effectiveness of the GDS constraint in unsupervised ReID networks is validated.
- Learned Block-Based Hybrid Image Compression74
This paper introduces explicit intra prediction into a learned image compression framework to utilize the relation among adjacent blocks and proposes a contextual prediction module (CPM) to better capture long-range correlations by utilizing the strip pooling to extract the most relevant information in neighboring latent space, thus achieving effective information prediction.
- Asynchronous Episodic Deep Deterministic Policy Gradient: Toward Continuous Control in Computationally Complex Environments74
This article proposes asynchronous episodic DDPG (AE-DDPG), a modified scheme for data collection in an asynchronous fashion, which can achieve more effective learning with less training time required and achieves higher rewards and two-fold to four-fold improvement in sample efficiency on average.
- ImmerTai: Immersive Motion Learning in VR Environments74
The results show that ImmerTai can accelerate the learning process of students noticeably compared to non-immersive learning with the conventional PC setup, and there is a substantial difference in the quality of the learnt motion between CAVE and HMD compared to PC.
- TuiGAN: Learning Versatile Image-to-Image Translation with Two Unpaired Images71
This paper proposes TuiGAN, a generative model that is trained on only two unpaired images and amounts to one-shot unsupervised learning that is capable of achieving comparable performance with the state-of-the-art UI2I models trained with sufficient data.
- KVQ: Kwai Video Quality Assessment for Short-form Videos69
The first short-form video quality evaluator, i.e., KSVQE, is proposed, which enables the quality evaluator to identify the quality-determined semantics with the content understanding of large vision language models (i.e., CLIP) and distinguish the distortions with the distortion understanding module.
- SeD: Semantic-Aware Discriminator for Image Super-Resolution68
The proposed Semantic-aware Discriminator (denoted as SeD) encourages the SR network to learn the fine-grained distributions by introducing the semantics of images as a condition, and empowered the SR network to produce more photo-realistic and pleasing images.
- View Invariant 3D Human Pose Estimation63
A view-invariant 3D human pose estimation module to alleviate the effects of viewpoint diversity and shows robustness for different 3D pose base networks and achieves a significant improvement over a powerful baseline on the public3D pose estimation benchmark Human3.6M.
- 63
- 62
- Stereoscopic Omnidirectional Image Quality Assessment Based on Predictive Coding Theory57
This article proposes a stereoscopic omnidirectional image quality evaluator (SOIQE) to cope with the characteristics of 3D 360-degree images and is a parametric model without necessary of regression learning, which ensures its interpretability and generalization performance.
- Exploring Explicit Domain Supervision for Latent Space Disentanglement in Unpaired Image-to-Image Translation56
An unpaired image-to-image translation framework, Domain-supervised GAN (DosGAN), which takes the first step towards the exploration of explicit domain supervision and pre-train a classification network to explicitly classify the domain of an image.
- 55
- 55
- 54
- Task-Driven Semantic Coding via Reinforcement Learning54
This work designs semantic maps for different tasks to extract the pixelwise semantic fidelity for videos/images and implements task-driven semantic coding by implementing semantic bit allocation based on reinforcement learning (RL).
- Learning Omni-Frequency Region-adaptive Representations for Real Image Super-Resolution53
This paper designs a Frequency Decomposition module to separate different frequency components to comprehensively compensate the information lost for real LR image and designs a Region-adaptive Frequency Aggregation module by leveraging dynamic convolution and spatial attention to adaptively restore frequency components for different regions.
- Perceptual Quality Assessment of Internet Videos52
A spatio-temporal distortion-aware model based on the properties of Internet videos originated from Youku, which outperforms the state-of-the-art methods and exhibits promising generalization ability is proposed.
- 52
- AI-GAN: Asynchronous interactive generative adversarial network for single image rain removal51
An Asynchronous Interactive Generative Adversarial Network (AI-GAN) is proposed to progressively disentangle the rainy image into background and rain spaces in feature level through a two-branch structure that outperforms state-of-the-art deraining methods and benefits various typical multimedia applications.
- Contralateral delay activity tracks object identity information in visual short term memory51
The results support the statement that CDA tracks the object identity information in VSTM as there was no difference on CDA between one color and four identical colors yet both were lower than 4 distinct colors.
- Image Super-Resolution Quality Assessment: Structural Fidelity Versus Statistical Naturalness50
This paper assesses the quality of SISR generated images in a two-dimensional (2D) space of structural fidelity versus statistical naturalness, and finds that a simple linear combination of a straightforward local structural fidelity and a global statisticalNaturalness measures produce surprisingly accurate predictions of S ISR image quality when tested using public subject-rated SISr image datasets.
- Learning Disentangled Feature Representation for Hybrid-Distorted Image Restoration50
This work introduces the concept of Disentangled Feature Learning to achieve the feature-level divide-and-conquer of hybrid distortions and proposes the feature disentanglement module (FDM) to distribute feature representations of different distortions into different channels by revising gain-control-based normalization.
- Adaptive Hypergraph Convolutional Network for No-Reference 360-degree Image Quality Assessment49
An adaptive hypergraph convolutional network for NR 360IQA, denoted as AHGCN, which first design a multi-level viewport descriptor for extracting hierarchical representations from viewports, and model interactions between viewports through hypergraphs, where each hyperedge connects two or more viewports.
- 49
- 48
- 47
- Light Field Image Quality Assessment With Auxiliary Learning Based on Depthwise and Anglewise Separable Convolutions46
This paper proposes an improved no-reference light field image quality assessment (NR-LFIQA) metric for future immersive media broadcasting services, and defines the spatial and angular feature estimations as auxiliary tasks in aiding the primary NR-L FIQA task by providing spatial andangular quality features as hints.
- 46
- 45
- 45
- 45
- Knowledge Graph-Based Image Classification Refinement44
This work proposes the concept of the image knowledge graph (IKG), a reasoning model of the knowledge graph that is closer to the biological visual information-processing model and conducts extensive experiments on large-scale image datasets (ImageNet), demonstrating the effectiveness of this approach.
- 42
- 40
- Learning based Facial Image Compression with semantic fidelity metric40
A Learning based Facial Image Compression framework with a novel Regionally Adaptive Pooling module whose parameters can be automatically optimized according to gradient feedback from an integrated hybrid semantic fidelity metric, including a successfully exploration to apply Generative Adversarial Network (GAN) as metric directly in image compression scheme.
- Region-Aware Text-to-Image Generation via Hard Binding and Soft Refinement39
RAG is a Regional-Aware text-to-image Generation method conditioned on regional descriptions for precise layout composition that achieves superior performance over attribute binding and object relationship than previous tuning-free methods.
- PromptCIR: Blind Compressed Image Restoration with Prompt Learning39
A prompt-learning-based compressed image restoration network, dubbed PromptCIR, which can effectively restore images from various compress levels, and exploits prompts to encode compression information implicitly, where prompts directly interact with soft weights generated from image features, thus providing dynamic content-aware and distortion-aware guidance for the restoration process.
- 3-D Context Entropy Model for Improved Practical Image Compression39
A 3-D context entropy model which can take advantage of known latent representation in current spatial locations for better entropy estimation is proposed and a light-weighted residual structure is adopted for feature learning during entropy estimation.
- Compression with Bayesian Implicit Neural Representations38
This work proposes overfitting variational Bayesian neural networks to the data and compressing an approximate posterior weight sample using relative entropy coding instead of quantizing and entropy coding it, which enables direct optimization of the rate-distortion performance by minimizing the $\beta$-ELBO.
- Uncertainty-Aware Few-Shot Image Classification38
This work proposes Uncertainty-Aware Few-Shot framework for image classification by modeling uncertainty of the similarities of query-support pairs and performing uncertainty-aware optimization and achieves the state-of-the-art performance.
- 38
- 36
- 35
- 35
- 35
- 35
- Full Reference Quality Assessment for Image Retargeting Based on Natural Scene Statistics Modeling and Bi-Directional Saliency Similarity35
The effectiveness of the BNSSD model is verified on two widely recognized public databases, and the experimental results show that the method outperforms the state-of-the-art algorithms under different statistical assessment criteria.
- Learning Cross-Scale Weighted Prediction for Efficient Neural Video Compression34
A novel cross-scale prediction module is proposed that achieves more effective motion compensation and a multi-stage quantization strategy is proposed, which improves the rate-distortion performance with no extra computational penalty during inference.
- SwinIQA: Learned Swin Distance for Compressed Image Quality Assessment34
A full-reference image quality assessment metric SwinIQA is designed to measure the perceptual quality of compressed images in a learned Swin distance space and achieves higher consistency with human’s perceptual judgment compared with both traditional methods and learning-based methods on CLIC datasets.
- Learning Distortion Invariant Representation for Image Restoration from a Causality Perspective33
This paper is the first to propose a novel training strategy for image restoration from the causality perspective, to improve the generalization ability of DNNs for unknown degradations.
- Lidar-Based 3D Obstacle Detection Using Focal Voxel R-CNN for Farmland Environment33
This study attempted to apply deep learning in LiDAR-based 3D obstacle detection for the farmland environment with significantly improves the detection performance for small objects, and the AP in the pedestrian class increased.
- FAN: Frequency Aggregation Network for Real Image Super-Resolution32
This work proposes FAN, a frequency aggregation network, to address the real-world image super-resolu-tion problem by extracting different frequencies of the LR image and passing them to a channel attention-grouped residual dense network (CA-GRDB) individually to output corresponding feature maps.
- 32
- Subjective Quality Assessment of Stereoscopic Omnidirectional Image31
This paper establishes the very first Stereoscopic OmnidirectionaL Image quality assessment Database (SOLID), and results demonstrate that the objective overall QoE assessment is more challenging compared to IQA in terms of stereoscopic omnidirectional images.
- 30
- Learned Video Compression with Feature-level Residuals30
This paper presents an end-to-end video compression network for P-frame challenge on CLIC, and combines the advantages of both pixel-level and feature-level residual compression methods by model ensembling.
- ZstGAN: An adversarial approach for Unsupervised Zero-Shot Image-to-image Translation29
This work proposes a framework called ZstGAN, which learns to model each domain with domain-specific feature distribution that is semantically consistent on vision and attribute modalities, and shows significant accuracy improvements over state-of-the-art zero-shot learning methods on CUB and FLO.
- Sequential Reinforced 360-Degree Video Adaptive Streaming With Cross-User Attentive Network29
A sequential RL-based (360SRL) ABR approach is proposed, transforming action space size of each decision step from exponential to linear via introducing a sequential decision structure, and results demonstrate that CUAN and 360SRL outperform existing viewpoint prediction and ABR approaches with a noticeable margin.
- Quality Assessment of Stereoscopic 360-degree Images from Multi-viewports29
A multi-viewport based full-reference stereo 360 IQA model is proposed that achieves a significant improvement over some well-known IQA metrics and can accurately reflect the overall QoE of perceived images.
- 29
- Active Token Mixer28
This work proposes an innovative token-mixer, dubbed Active Token Mixer (ATM), to actively incorporate contextual information from other tokens in the global scope into the given query token, taking ATMs as the primary operators and assembling them into a cascade architecture, dubbed ATMNet.
- Semantically Structured Image Compression via Irregular Group-Based Decoupling28
This paper proposes the concept of group-independent transform that maintain the independence among distinct groups that decouple an image into multiple groups with irregular shapes based on a customized group mask and compress them independently.
- SASL: Saliency-Adaptive Sparsity Learning for Neural Network Acceleration27
A Saliency-Adaptive Sparsity Learning (SASL) approach for further optimization of CNNs, where the regularization strength is adjusted according to the saliency, so the optimized format can better preserve the prediction performance while zeroing out more computation-heavy filters.
- LIRA: Lifelong Image Restoration from Unknown Blended Distortions27
Experimental results show that the proposed approach can not only achieve state-of-the-art performance on blended distortions removal tasks in both PSNR/SSIM metrics, but also maintain old expertise while learning new restoration tasks.
- 27
- DeepIR: A Deep Semantics Driven Framework for Image Retargeting27
This work presents Deep Image Retargeting (DeepIR), a coarse-to-fine framework for content-aware image retargeting that first constructs the semantic structure of input image with a deep convolutional neural network and a uniform re-sampling is devised to resize feature maps to target aspect ratio at each feature layer.
- 27
- 27
- 27
- Ntire 2025 Challenge on Short-Form Ugc Video Quality Assessment and Enhancement: Methods and Results26
This paper presents a review for the NTIRE 2025 Challenge on Short-form UGC Video Quality Assessment and Enhancement, and introduces a new short-form UGC dataset tailored for single-image super-resolution, i.e., the KwaiSR dataset.
- StyleAM: Perception-Oriented Unsupervised Domain Adaption for No-Reference Image Quality Assessment26
This work proposes an effective perception-oriented unsupervised domain adaptation method StyleAM (Style Alignment and Mixup) for NR-IQA, which transfers sufficient knowledge from label-rich source domain data to label-free target domain images.
- Deep Multi-Scale Features Learning for Distorted Image Quality Assessment26
This model is based on both residual maps and distorted images in luminance domain, where the proposed network contains spatial pyramid pooling and feature pyramid from the network structure, and is optimized in a deep end-to-end supervision manner.
- Learned Scalable Image Compression with Bidirectional Context Disentanglement Network26
A learned scalable/progressive image compression scheme based on deep neural networks (DNN), named Bidirectional Context Disentanglement Network (BCD-Net), which outperforms the state-of-the-art DNN-based scalable image compression methods in both PSNR and MS-SSIM metrics.
- Recent Advances on HEVC Inter-Frame Coding: From Optimization to Implementation and Beyond26
This paper provides a comprehensive review of the state-of-the-art techniques for HEVC inter-frame coding from three aspects, namely fast inter coding solutions, implementation on different hardware platforms as well as advanced inter coding techniques.
- Integrating seismic first-break picking methods with a machine learning approach26
This study builds an architecture of convolutional neural networks in machine learning, helping identify poor picks across multiple traces from seismic data, and applies a machine learning method to identify and fix poor picks automatically.
- 26
- 26
- 25
- NVTC: Nonlinear Vector Transform Coding25
This paper proposes a novel framework for neural image compression named Nonlinear Vector Transform Coding (NVTC), which solves the critical complexity issue of VQ through a multi-stage quantization strategy and nonlinear vector transforms.
- Guest Editorial QoE-Aware Wireless Multimedia Systems25
The 11 papers in this special issue cover a range of topics and can be logically organized in three groups, focusing on QoE-aware media protection, QOE assessment and modelling, and multi-user-QoE management.
- Bayesian graph convolutional network for traffic prediction24
A Bayesian Graph Convolutional Network (BGCN) framework is proposed to alleviate issues of spatial relationships between traffic conditions due to ignoring the prior of the observed topology of the road network, and neglecting the presence of negative spatial relationships.
- Is Vanilla MLP in Neural Radiance Field Enough for Few-Shot View Synthesis?24
This paper proposes the multi-input MLP (mi-MLP) that incorpo-rates the inputs of the vanilla MLP into each layer to prevent the overfitting issue without harming detailed synthesis and proposes to model colors and volume density separately and present two regularization terms.
- Multi-scale Grouped Dense Network for VVC Intra Coding24
The multi-scale grouped dense network (MSGDN) is designed to further reduce the compression artifacts by combining the multi- scale and grouped dense block, which are integrated as the post-process network of VVC intra coding.
- Prompt-ICM: A Unified Framework towards Image Coding for Machines with Task-driven Prompts23
Inspired by recent advances in transferring large-scale pre-trained models to downstream tasks via prompting, this work explores a new ICM framework, termed Prompt-ICM, which could efficiently support different kinds of intelligent tasks with much higher coding efficiency.
- No-Reference Light Field Image Quality Assessment Based on Micro-Lens Image23
This paper proposes a No-Reference Light Field image Quality assessment model based on MLI (LF-QMLI), which utilizes Global Entropy Distribution and Uniform Local Binary Pattern descriptor (ULBP) to extract features from the MLI, and pool them together to measure angular consistency.
- Improved image classification with 4D light-field and interleaved convolutional neural network23
A new deep convolutional neural network is proposed to improve image classification using extra light-field angular information by replacing the fully connected layer of a VGG network with a set of interleaved spatial-angular filters, thus providing more accurate classification performance over traditional models.
- GaussianSR: 3D Gaussian Super-Resolution with 2D Diffusion Priors22
This work proposes to leverage off-the-shelf 2D diffusion priors by distilling the 2D knowledge into 3D with Score Distillation Sampling (SDS), and introduces two simple yet effective techniques to reduce stochastic disturbances introduced by SDS.
- 22
- Texture-Enhanced Light Field Super-Resolution With Spatio-Angular Decomposition Kernels22
This article comprehensively discover the potentials of LF decomposition, systematically unify the decomposition operations of various subspaces into a series of decomposition kernels, which are incorporated into the proposed decomposition kernel network (DKNet) for comprehensive spatio-angular feature extraction.
- 22
- Re-energizing Domain Discriminator with Sample Relabeling for Adversarial Domain Adaptation22
This work proposes an efficient optimization strategy named Re-enforceable Adversarial Domain Adaptation (RADA), which aims to re-energize the domain discriminator during the training by using dynamic domain labels, and relabeling makes the less separable distributions more separable, and thus leads to a more powerful domain classifier w.r.t. the new data distributions.
- JPEG Pleno: a standard framework for representing and signaling plenoptic modalities22
An overview of a new standardization initiative – JPEG Pleno – that is intended to define an efficient framework addressing the above interoperability issues is provided about its current status and future plans.
- 21
- 21
- 21
- Beyond Triplet Loss: Meta Prototypical N-Tuple Loss for Person Re-identification20
A multi-class classification loss, i.e., N-tuple loss, is introduced to jointly consider multiple ($N$) instances for per-query optimization for person ReID, and achieves the state-of-the-art performance on the benchmark person Re ID datasets.
- 20
- 20
- Ordering graphs with small index and its application20
The asymptotic ordering for the connected graphs with index less than √2+√5 is determined and its application to the study of acyclic Kekulean molecules with big HOMO-LUMO separation is given.
- Hybrid Agents for Image Restoration19
This work presents HybridAgent, intending to incorporate multiple restoration modes into a unified image restoration model and achieve intelligent and efficient user interaction through the authors' proposed hybrid agents, and proposes the hybrid rule of fast, slow, and feedback restoration agents.
- OmniV2V: Versatile Video Generation and Editing via Dynamic Content Manipulation19
OmniV2V is proposed, a video model capable of generating and editing videos across different scenarios based on various operations, including: object movement, object addition, mask-guided video edit, try-on, inpainting, outpainting, human animation, and controllable character video synthesis.
- 19
- VisualCritic: Making LMMs Perceive Visual Quality Like Humans19
This paper presents VisualCritic, the first LMM for broad-spectrum image subjective quality assessment, which enables new capabilities of quantitatively measuring the perceptual quality of given images in terms of their Mean Opinion Score, noisiness, colorfulness, sharpness, and other numerical indicators.
- Deep Local and Global Spatiotemporal Feature Aggregation for Blind Video Quality Assessment19
An efficient VQA method named Deep SpatioTemporal video Quality assessor (DeepSTQ) to predict the perceptual quality of various distorted videos in a no-reference manner and outperforms state-of-the-art quality assessment algorithms.
- Image-to-Image Translation with Multi-Path Consistency Regularization19
A new kind of loss, multi-path consistency loss, is introduced, which evaluates the differences between direct translation $\mathcal{D}_s\to\mathcal(D)_t$ and indirect translation, to regularize training.
- 19
- 19
- Sequential Gating Ensemble Network for Noise Robust Multiscale Face Restoration18
Experimental results demonstrate that the proposed sequential gating ensemble network (SGEN) is more effective at multiscale human face restoration with more image details and less noise than state-of-the-art image restoration models.
- Stereoscopic Video Quality Prediction Based on End-to-End Dual Stream Deep Neural Networks18
A no-reference stereoscopic video quality assessment (NR-SVQA) method based on an end-to-end dual stream deep neural network (DNN), which incorporates left and right view sub-networks, which outperforms state-of-the-art algorithms.
- Local Patch AutoAugment With Multi-Agent Collaboration17
This paper proposes a more fine-grained automated DA approach, dubbed Patch AutoAugment, to divide an image into a grid of patches and search for the joint optimal augmentation policies for the patches as a multi-agent reinforcement learning (MARL) problem.
- Automatic microseismic detection and location via the deep-convolutional neural network17
This work constructs an automatic processing scheme by combining the detection and location neural networks for the continuous records and applies the method into the field datasets, showing that the time window of a complete event can be detected effectively even though the CNN model is trained with a small number of samples.
- GSemSplat: Generalizable Semantic 3D Gaussian Splatting from Uncalibrated Image Pairs16
GSemSplat is introduced, a framework that learns open-vocabulary semantic representations linked to 3D Gaussians without the need for per-scene optimization, dense image collections or calibration, and employs a dual-feature approach that leverages both region-specific and context-aware semantic features as supervision in the 2D space.
- Meta Clustering Learning for Large-scale Unsupervised Person Re-identification16
This paper makes the first attempt to the large-scale U-ReID and proposes a "small data for big task" paradigm dubbed Meta Clustering Learning (MCL), which only pseudo-labels a subset of the entire unlabeled data via clustering to save computing for the first-phase training.
- 16
- Reinforced Bit Allocation under Task-Driven Semantic Distortion Metrics16
This paper formulate the bit allocation problem as a Markovian Decision Process (MDP) and train RL agents to automatically decide the quantization parameter (QP) of each coding tree unit (CTU) for HEVC intra coding, according to the task-driven semantic distortion metrics.
- Visual Comfort Assessment for Stereoscopic Image Retargeting16
A Visual Comfort Assessment metric for Stereoscopic Image Retargeting (VCA-SIR) is proposed, which introduces novel features like disparity range, boundary disparity as well as disparity intensity distribution into the assessment model.
- Efficient VR Video Representation and Quality Assessment16
The experimental results show that the proposed Quadrangle Affine Square Projection (QASP) method achieves noticeable improvement over traditional methods, and the proposed quality assessment method outperforms the traditional measurements in terms of consistency with subjective evaluation.
- Multi-Scale Face Restoration With Sequential Gating Ensemble Network16
The principle of ensemble learning is employed into SGEN architecture design to reinforce predictive performance of the network and demonstrates that the SGEN is more effective at multi-scale human face restoration with more image details and less noise than state-of-the-art image restoration models.
- 16
- SelectAugment: Hierarchical Deterministic Sample Selection for Data Augmentation15
This paper proposes an effective approach to select samples for augmentation in a deterministic and online manner based on the sample contents and the network training status, and adopts Hierarchical Reinforcement Learning (HRL) to learn the selection policy.
- Source-free Unsupervised Domain Adaptation for Blind Image Quality Assessment15
This paper takes the first step towards the source-free unsupervised domain adaptation (SFUDA) in a simple yet efficient manner for BIQA to tackle the domain shift without access to the source data.
- Light field reconstruction using hierarchical features fusion15
A novel end-to-end U-Net with SAS network (U-SAS-Net) is introduced to extract and fuse hierarchical features, both local and semantic, from a relatively large receptive field while establishing the relation of the correlated sub-aperture images.
- 15
- Oriented total variation l1/2 regularization15
It is theoretically and experimentally demonstrated that applying l1/2 regularization to the directional derivatives of images leads to superior structure preservation, and the experimental results verify that OTV l 1/2 outperforms other similar models.
- A packet-layer video quality assessment model with spatiotemporal complexity estimation15
This article proposes a novel video quality monitoring model that takes into account the spatiotemporal complexity of video content, and achieves robust performance improvement compared with the existing peer VQA metrics on both dataset-1 and dataset-2.
- 14
- 14
- 14
- Integral sum graphs from identification14
The methods of identification are introduced and it is proved that the generalized stars and the trees with all forks at least distance 4 apart are integral sum graphs.
- Versatile Neural Processes for Learning Implicit Neural Representations13
This paper introduces a bottleneck encoder that produces fewer and informative context tokens, relieving the high computational cost while providing high modeling capability, and proposes an efficient NP framework dubbed Versatile Neural Processes (VNP), which largely increases the capability of approximating functions.
- 13
- Multi-task Learning-based All-in-one Collaboration Framework for Degraded Image Super-resolution13
Extensive experimental results on synthetic and real-world low-resolution images show that the proposed all-in-one collaboration framework not only produces favorable high-resolution results while removing serious degradation, but also has high computational efficiency, outperforming state-of-the-art methods.
- 13
- 12
- Survey on Visual Signal Coding and Processing With Generative Models: Technologies, Standards, and Optimization12
This paper provides a survey of the latest developments in visual signal coding and processing with generative models, and presents the latest advancements in fast optimization on visual signal coding and processing with generative models.
- FreqAlign: Excavating Perception-Oriented Transferability for Blind Image Quality Assessment From a Frequency Perspective12
This work proposes an effective alignment strategy, i.e., Frequency Alignment (dubbed FreqAlign), to excavate the perception-oriented transferability of BIQA in the frequency space by performing feature frequency decomposition and selecting the frequency components that contained the most transferable perception knowledge for alignment.
- Learning to Transfer: Unsupervised Domain Translation via Meta-Learning12
This work proposes a model called Meta-Translation GAN (MT-GAN) to find good initialization of translation models and demonstrates effectiveness of the model on ten diverse two-domain translation tasks and multiple face identity translation tasks.
- 12
- 12
- 11
- 11
- 11
- 11
- Relation-Aware Global Attention.11
This paper proposes an effective Relation-Aware Global Attention module for CNNs to fully exploit the global correlations to infer the attention and demonstrates the general applicability of RGA to vision tasks by applying it to the scene segmentation and image classification tasks resulting in consistent performance improvement.
- 11
- Exploiting repeated patterns for efficient compression of massive models11
An efficient compression algorithm for massive models, which consist of a large number of small to medium sized connected components, based on efficiently exploiting repetitive patterns in the input model is proposed, which can save 40% storage on average in comparison with gzip, the popular compression software used in [Shikhare et al. 2001].
- Exploration and visualization of OLAP cubes with statistical tests11
A two-dimensional checkerboard visualization of the cube is introduced that allows interactive exploration to understand significant measure differences between two cuboids differing in one dimension along with associated image data.
- 11
- 11
- Tell Codec What Worth Compressing: Semantically Disentangled Image Coding for Machine with LMMs10
A new image compression paradigm to achieve "intelligently coding for machine" by cleverly leveraging the common sense of Large Multimodal Models, called SDComp for "Semantically Disentangled Compression", which leads to more flexible reconstruction results, promised decoded visual quality, and a more generic/satisfactory intelligent task-supporting ability.
- QMamba: On First Exploration of Vision Mamba for Image Quality Assessment10
This work takes the first exploration of the recently popular foundation model, i.e., State Space Model/Mamba, in image quality assessment (IQA), aiming at observing and excavating the perception potential in vision Mamba, and proposes the StylePrompt tuning paradigm, which enables better perceptual transfer with lower computational cost.
- 10
- 10
- 10
- 10
- A Decomposed Dual-Cross Generative Adversarial Network for Image Rain Removal.10
The proposed Decomposed Dual-Cross Generative Adversarial Network (DDC-GAN) shows performance improvement compared with state-of-the-art methods on both synthetic and real-world images in terms of qualitative and quantitative measures.
- 10
- 10
- Forcing faces in plane bipartite graphs10
It is proved that any connected plane bipartite graph with a forcing face is elementary, and it is shown that for any integers n and k, there exists a plane elementary bipartites graph such that exactly k of the n finite faces of G are forcing.
- On Integral Sum Graphs10
The structure for all integral sum graphs with exactly two saturated nodes are determined, and an upper bound for the number of edges of a connected integral sum graph with no saturated nodes is given.
- AdaNav: Adaptive Reasoning with Uncertainty for Vision-Language Navigation9
AdaNav is proposed, an uncertainty-based adaptive reasoning framework for VLN that introduces Action Entropy as a policy prior for UAR and progressively refine it through a Heuristics to RL training method, enabling agents to learn difficulty aware reasoning policies under the strict data limitations of embodied tasks.
- 9
- 9
- 9
- An azimuthal Fourier domain formulation of the Ffowcs Williams and Hawkings equation9
The Ffowcs Williams and Hawkings equation is manipulated to obtain a new formulation that lives entirely in the azimuthal Fourier domain and can be used to efficiently and naturally compute the acoustic field associated with resolvent modes of a Mach 1.5 jet, thus avoiding the need to compute the modes on a large computational domain to capture their acoustic radiation.
- Hierarchical visual comfort assessment for stereoscopic image retargeting9
A Hierarchical Visual Comfort Assessment (Hi-VCA) scheme for SIR, considering hybrid distortions including structure, information, semantic distortions usually occurring in retargeting, and binocular incongruity existing in stereoscopic multimedia is proposed.
- 9
- 9
- OLAP-based query recommendation9
This paper developed an online OLAP algorithm to generate query suggestions for the users based on the frequency of the keywords in the selected documents and the correlation between the words in the collection, and shows valid recommendations as a result of combinations generated using the correlations between the keywords.
- On k-pairable graphs9
The concept of k-pairable graphs is introduced and a new graph parameter p(G), called the pair length of graph G, is defined, which measures the maximum distance, in some sense, between a subgraph induced by half the vertices of G with the isomorphic sub graph induced by the other half of V(G).
- Semantic-Aware Message Broadcasting for Efficient Unsupervised Domain Adaptation8
A novel method, Semantic-aware Message Broadcasting (SAMB), which enables more informative and flexible feature alignment for unsupervised domain adaptation (UDA) and systematically study the effects of adversarial-based feature alignment (ADA) and pseudo-label based self-training (PST) on UDA.
- 8
- 8
- Deliberation Learning for Image-to-Image Translation8
Inspired by the success of deliberation network in natural language processing, deliberation process is extended to the field of image translation by considering an additional polishing step on the output image.
- CMC: Few-shot Novel View Synthesis via Cross-view Multiplane Consistency7
This work proposes a simple yet effective method that explicitly builds depth-aware consistency across input views and can achieve better synthesis quality over state-of-the-art methods.
- FPGA-Based Improved Sobel Operator Edge Detection7
Compared to other edge detection algorithms, the FPGA-based improved Sobel operator edge detection method efficiently and rapidly detects image edges, ensuring stable image display and accurate edge information.
- 7
- 7
- 7
- 7
- CASINet: Content-Adaptive Scale Interaction Networks for scene parsing7
A Content-Adaptive Scale Interaction Network (CASINet) to exploit the multi-scale features for scene parsing based on the classic Atrous Spatial Pyramid Pooling module, followed by a proposed contextual scale interaction (CSI) module, and a scale adaptation (SA) module.
- 7
- A Coarse-to-Fine Framework for Learned Color Enhancement with Non-Local Attention7
This paper proposes a coarse-to-fine framework with non-local attention for color enhancement that divides enhancement process into channel-wise enhancement and pixel-wise refinement performed by two cascaded Convolutional Neural Networks (CNNs).
- 7
- 7
- 6
- Hierarchical Reinforcement Learning Based Video Semantic Coding for Segmentation6
This paper proposes the Hierarchical Reinforcement Learning based task-driven Video Semantic Coding, named as HRLVSC, and divides the action space into frame-level and CTU-level spaces in a hierarchical manner, and designs a simple but effective mode simplification strategy for it.
- Light Field Compression Based on Implicit Neural Representation6
This work proposes a novel light field compression scheme based on implicit neural representation to reduce redundancies between views and achieves comparable rate-distortion performance as well as superior perceptual quality over traditional methods.
- 6
- 6
- Predictive Auto-Encoding Network for Blind Stereoscopic Image Quality Assessment.6
A Predictive Auto-encoDing Network (PAD-Net) for blind/No-Reference stereoscopic image quality assessment (NR-SIQA) is proposed and the Siamese framework is designed to mimic the binocular rivalry in the 3D HVS based on the likelihood and prior maps generated from the predictive coding process.
- None Ghosting Artifacts Stitching Based on Depth Map for Light Field Image6
Compared with state-of-the-art stitching methods aiming to alleviate ghosting artifacts, the proposed None Ghosting Artifacts (NGA) stitching method not only stitches images without ghosted artifacts, but also achieves realistic perspective transformation.
- Seismic trace editing by applying machine learning6
This study proposes a method based on machine learning to conduct seismic trace editing automatically and combines the techniques of Hough transform and Convolutional Neural Network for trace editing.
- 6
- 6
- 6
- InternVQA: Advancing Compressed Video Quality Assessment with Distilling Large Foundation Model5
A distillation method to equip the smaller model with rich compression quality priors is proposed and the results showed that the lightweight model distilled from InternVideo2 achieved excellent performance in compression video quality assessment.
- Learned Lossless JPEG Transcoding via Joint Lossy and Residual Compression5
This work is the first to utilize the learned end-to-end lossy transform coding to reduce the redundancy of DCT coefficients in a compact representational domain and introduces residual compression for lossless transcoding, which adaptively learns the distribution of residual DCT coefficient before compressing them using context-based entropy coding.
- 5
- Rethinking Classification Loss Designs for Person Re-identification with a Unified View.5
It is argued that triplet-based optimization can be viewed as a two-class subsampling classification, which performs classification over two sampled categories based on instance similarities, and a new meta prototypical N-tuple loss is proposed for more efficient multi- class subsampled classification.
- 5
- 5
- 5
- Augmented Coarse-to-Fine Video Frame Synthesis with Semantic Loss5
This paper introduces augmented inputs to compensate texture details and motion information and presents a coarse-to-fine guidance scheme with a well-designed semantic loss to improve the capability of video frame synthesis.
- ONTOCUBE5
This paper presents a novel approach based on using OLAP cubes for ontology extraction through an analytical process of the summarized frequencies of keywords within a corpus and shows how all the proposed discrimination measures affect the resulting classes.
- A packet-layer video quality assessment model based on spatiotemporal complexity estimation5
A novel packet-layer video quality assessment model dealing with video quality monitoring is proposed by incorporating the estimation of the spatiotemporal complexity of video content and achieves preferable and robust performance improvement compared with the existing models.
- DBDOC: querying and browsing databases and interrelated documents5
This work developed a system based on SQL code generation and User-Defined Functions that analyzes document-to-metadata links by extracting a basic set of relationships at different levels of granularities: coarse, medium and fine, and provides typical information retrieval capabilities for querying medium-sized document collections of interrelated documents in the DBMS.
- 5
- Adaptive Video Presentation for Small Display While Maximize Visual Information5
An intact and generic framework to adapt video presentation (AVP) is presented and a novel method for choosing the optimal cropped region is introduced to minimize the information loss over adapting video presentation.
- Revisiting MLLM Token Technology through the Lens of Classical Visual Coding4
This study presents the first comprehensive and structured technology comparison of MLLM token and visual coding, paving the way for more efficient multimodal models and more powerful visual codecs simultaneously.
- TIV-Diffusion: Towards Object-Centric Movement for Text-driven Image to Video Generation4
This work proposes a new diffusion-based TI2V framework, termed TIV-Diffusion, intending to achieve precise control and high-quality video generation based on textual-described motion for different objects, and introduces an object-centric textual-visual alignment module.
- LiftVSR: Lifting Image Diffusion to Video Super-Resolution via Hybrid Temporal Modeling with Only 4$\times$RTX 4090s4
LiftVSR is proposed, an efficient VSR framework that leverages and elevates the image-wise diffusion prior from PixArt, achieving state-of-the-art results using only 4$\times$RTX 4090 GPUs and introduces a hybrid temporal modeling mechanism that decomposes temporal learning into two complementary components.
- Research on Segmentation Technology in Lung Cancer Radiotherapy Based on Deep Learning4
The advancement of deep learning-based segmentation technology for lung cancer radiotherapy (DSLC) research was examined from the perspectives of LTs and OARs and the issues that DSLC faces are discussed, as well as possible viable solutions.
- 4
- 4
- 4
- 4
- 4
- 4
- 4
- Raster image representation of fingerprint minutiae4
The proposed method enhances the privacy of a fingerprint minutiae template against information leakage and applies to any other cases requiring a transformation of discrete points based features to fixed-length and ordered features, e.g., media fingerprinting/hashing applications.
- Compression of massive models by efficiently exploiting repeated patterns4
A new compression algorithm for massive models, which consist of a large number of small to medium sized connected components, is proposed, which is more efficient on detecting repeated components by recognizing instances repeating in various scalings.
- 4
- 4
- Denoising Saliency Map for Region of Interest Extraction4
The filter presented in this paper makes use of the property of saliency maps and can remove almost all the Gauss noise and pepper-salt noise while preserve the details of attention area.
- 4
- Quality Weighted Bit Allocation for Smoother Streaming of Stored FGS Video4
A Fine Granularity Scalability streaming framework together with a Quality Weighted Bit Allocation scheme for enhancement-layer is first proposed to achieve smooth video quality even under channel conditions with a wide range of bandwidth variation.
- 3
- 3
- 3
- UniMIC: Towards Universal Multi-modality Perceptual Image Compression3
The UniMIC introduces the visual codec repository, which incorporates amounts of representative image codecs and directly uses them as the basic codecs for various practical applications and proposes multi-grained textual coding, where variable-length content prompt and compression prompt are designed and encoded to assist the perceptual reconstruction through the multi-modality conditional generation.
- 3
- 3
- Diagnosis-oriented Medical Image Compression with Efficient Transfer Learning3
This study proposes DMIC, the first efficient transfer learning-based codec for diagnosis-oriented medical image compression, which can be effectively optimized with only few-shot annotated medical examples, by reusing the knowledge in the existing reinforcement learning- based task-driven semantic coding framework, i.e., HRLVSC.
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- GTA: Advancing Image-to-3D World Generation via Geometry Then Appearance Video Diffusion2
GTA is shown to be effective as a general enhancement module that further improves the generation quality of existing image-to-3D world pipelines, as well as supporting multiple downstream applications and exhibiting favorable data efficiency during model training, highlighting its versatility and broad applicability.
- Why Compress what you can Generate? When GPT-4o Generation Ushers in Image Compression Fields2
This work decouple the image signal into textual descriptions and a low-resolution visual prior and investigates two typical compression paradigms: textual coding and multimodal coding, where all/most pixel-level information is generated instead of compressing via the advanced GPT-40 image generation function.
- A close look at few-shot real image super-resolution from the distortion relation perspective2
This work is the first to have a close look at the under-explored direction for RealSR, i.e., few-shot real-world image super-resolution, and proposes Distortion Relation guided Transfer Learning (DRTL) for the few- shot RealSR by transferring the rich restoration knowledge from auxiliary distortions to the target RealSR under the guidance of distortion relation.
- Balancing Accuracy and Computational Efficiency: A Faster R-CNN with Foreground-Background Segmentation-Based Spatial Attention Mechanism for Wild Plant Recognition2
Compared to the original Faster R-CNN, the improved lightweight Faster R-CNN architecture named ULS-FRCN is suitable for deployment on resource-constrained forestry devices, enabling efficient plant identification and management without the need for high-performance servers.
- 2
- A Logical Framework with Higher-Order Rational (Circular) Terms2
CoLF is proposed, a conservative extension of LF with higher-order rational terms and mixed inductive and coinductive definitions, which proves that two terms are equal if they unfold to the same infinite regular Bohm tree.
- StyleAM: Perception-Oriented Unsupervised Domain Adaption for Non-reference Image Quality Assessment2
- 2
- 2
- 2
- 2
- 2
- 2
- Unpaired Image-to-Image Translation with Domain Supervision.2
This work designs an unpaired image-to-image translation framework, Domain-supervised GAN (briefly, DosGAN), that takes the first step towards exploration of domain supervision, and transfers the domain feature extractor obtained on the Facescrub dataset to the CelebA dataset without domain supervision information.
- Research on Active Interaction Design for Smart Speakers Agent of Home Service Robot2
The active interaction design method is proposed to enhance the initiate of the intelligent agents to solve the user’s needs, improve Interaction performance and user experience and gradually improve the active interaction research of the home service robot in the family environment.
- 2
- 2
- 2
- 2
- 2
- 2
- 1
- MiNL: Micro-Images based Neural Representation for Light Fields1
MiNL is proposed, a novel MI-wise implicit neural representation for light fields that can transform the light field compression task into model compression task and can achieve comparable performance with state-of-the-art methods, with about 1~2dB PSNR improvement over HEVC/H.265 at the same bit rate.
- LossAgent: Towards Any Optimization Objectives for Image Processing with LLM Agents1
The LossAgent introduces the powerful large language model (LLM) as the loss agent, where the rich textual understanding of prior knowledge empowers the loss agent with the potential to understand complex optimization objectives, trajectory, and state feedback from external environments in the optimization process of the low-level image processing networks.
- 1
- 1
- 1
- A Logical Framework with Infinitary Terms1
CoLF is proposed as an alternative interpretation of CoLF-style signatures where terms are taken to be all possibly infinitary terms that are consistent with a given signature, and the notion of productive B\"ohm trees, a particular kind of typed $\bot$-free B \"ohm trees that are closed under hereditary substitution are proposed.
- 1
- 1
- 1
- 1
- 1
- 1
- Latent-Separated Global Prediction for Learned Image Compression.1
This paper proposes a 3-D global context model, which separates the latents into two channel groups, and will leverage the known group to model spatial correlations that guide the global prediction for the unknown group and thus achieve more efficient entropy estimation.
- Distribution Discrepancy Maximization for Image Privacy Preserving1
This work proposes a collaborative training scheme that theoretically proves that such a training scheme maximizes two distributions' discrepancy, and can produce satisfactory defense against the attack of deep recognition models indicated by significant accuracy decreases on Casia-WebFace and LFW datasets.
- 1
- 1
- Characterizing the near-surface velocity structures by applying machine learning1
A near-surface characterization method based on the Convolution Neural Network (CNN), a machine learning approach that automatically identifies several pre-defined model types from shot or receiver gathers, and it includes two steps: training and prediction.
- 1
- Study on User Experience of Panoramic Images on Different Immersive Devices1
This paper analyzes five factors: the sense of immersion, sense of involvement, comfort factor, resolution factor and content factor presented by three typical types of panoramic display devices: CAVE, Oculus rift, and web panorama player.
- 1
- 1
- Adaptive sample map for Monte Carlo ray tracing1
A novel noise artifact detection and noise level representation method is proposed and the result is adaptive sample map (ASP), which serves two purposes in rendering process; its statistics information can be used as noise level metric in synthetic image, and it can also be used by a Monte Carlo ray tracer to refine the synthetic image adaptively in order to reduce the noise to unnoticeable level with less rendering time than the brute force method.
- 1
- Pattern-based assembled DCT scheme for image coding1
PADCT has the advantages of advanced compact energy concentration, efficient usage of memory, fewer categories of one dimensional transform, and adaptability for hardware optimization and outperforms significantly than previous typical directional transform schemes in comparison with PSNR or subjective quality.
- 1
- –
- –
- –
- Semantics Disentanglement and Composition for Universal Image Coding With Efficiently LLM Reasoning and Generative Diffusion–
UniCodec, a universal codec built on a novel paradigm of semantic disentanglement at the encoder and compositional generation at the decoder, is introduced, effectively bridging the gap between human-centric and machine-centric compression.
- –
- –
- –
- Compliance With Physical Activity, Sedentary Behaviour and Sleep Recommendations in Children Aged 5–13 Years: A Systematic Review and Meta‐Analysis Across 15 Countries–
Overall adherence to the 24-h Movement Guidelines among school-aged children is low, with the lower compliance for MVPA&ST guideline and MVPA guideline, and targeted interventions to increase MVPA may improve overall compliance.
- Querying Graph-Relational Data–
The graph-relational database model is formally defined and a static and dynamic semantics for queries are presented, which provides a flexible, compositional, and strongly-typed solution to this"object-relational mismatch."
- –
- –
- Towards Defining an Efficient and Expandable File Format for AI-Generated Contents–
A new file format for AIGC images, named AIGIF, enabling ultra-low bitrate coding of AIGC images and introduces an expandable syntax in AIGIF to support the extension of the most advanced generation models to be developed in the future.
- –
- –
- –
- Research on the Path of Coordination and Optimization Between Information System and Business Strategy–
Research findings indicate that IS-business strategy synergy requires not only close technical-management integration but also support from corporate culture, organizational structure, and decision-making mechanisms, which means well-designed optimization pathways can better adapt to complex market environments, enhancing operational flexibility and competitiveness.
- –
- –
- Research on Image Defogging Algorithm based on FPGA–
Experimental results show that the image dehazing effect can be effectively improved by combining software and hardware.
- –
- –
- Review of Research on Feature Extraction Algorithms for Finger Vein Images–
The finger vein recognition system is introduced and organizes and summarizes the public datasets, and the applications of vein feature extraction in recent years are researched, and the status quo of the finger vein feature extraction is summed up.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- Detection of the near surface velocity reveal with the convolutional neural network–
This work develops a method based on the convolutional neural network (CNN) for automatically detecting the shingling features directly from 3D seismic data and predicts shingle features with high accuracy.
- –
- –
- –
- Discovering frequent pattern pairs–
This work introduces cube pairs and rule pairs, which are pattern pair generalizations of cubes and association rules, respectively, and introduces a reliability metric based on cross-validation, which enables an objective comparison between both patterns.
- Optimization and gereralization of olap cube processing in relational database systems–
By combining OLAP cubes with a parametric statistical test, it is shown that Cube Statistical Test is superior in terms of the patterns, and with OLAP as a data mining tool, it can synergistically work with statistics.
- Android Platform-based RSS Reading System Design and Implementation–
The architecture of Android system is introduced and the design and implementation of the Android platform-based RSS reader system is illustrated with the popular RSS technology and web services provided by Java Servlet.
- Robust discovery of partial rigid symmetries on 3D models–
This work states that automatic discovery of partial rigid symmetry on general 3D models, with no assumption on the size, shape or location of the symmetric parts, keeps to be a hot topic during recent years.
- –
- –
- –
- –
- Error Resilient Video Compression Using Behavior Models–
This paper study the behavior of a predictive (interframe) video encoder and model the encoders behavior using only the statistics of the original input data and of the underlying channel prone to bit errors to control the rate of the videoencoder.
- –
- –
- –
- –
- –
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.