Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks272 from public data
- Enhanced-alignment Measure for Binary Foreground Map Evaluation1,716
This paper takes a detailed look at current binary FM evaluation measures and proposes a novel and effective E-measure (Enhanced-alignment measure), jointly capturing image-level statistics and local pixel matching information, which demonstrates the superiority of this measure over the available measures on 4 popular datasets via 5 meta-measures.
- EGNet: Edge Guidance Network for Salient Object Detection1,202
This paper presents an edge guidance network (EGNet) for salient object detection with three steps to simultaneously model these two kinds of complementary information in a single network to solve the complementarity between salient edge information and salient object information.
- 412
- 328
- 254
- Enhancing text clustering by leveraging Wikipedia semantics223
A way to build a concept thesaurus based on the semantic relations (synonym, hypernym, and associative relation) extracted from Wikipedia is proposed and a unified framework to leverage these semantic relations in order to enhance traditional content similarity measure for text clustering is developed.
- Self-Sustaining Representation Expansion for Non-Exemplar Class-Incremental Learning212
A novel self-sustaining representation expansion scheme that fuses main-br branch expansion and side-branch updating to maintain the old features and transfer the invariant knowledge to the new classes is proposed.
- 201
- Self-Promoted Prototype Refinement for Few-Shot Class-Incremental Learning193
This work proposes a novel incremental prototype learning scheme which consists of a random episode selection strategy that adapts the feature representation to various generated incremental episodes to enhance the corresponding extensibility, and a self-promoted prototype refinement mechanism which strengthens the expression ability of the new classes by explicitly considering the dependencies among different classes.
- Nighttime Dehazing with a Synthetic Benchmark181
A novel synthetic method to simulate nighttime hazy images from daytime clear images, which first reconstructs the scene geometry, then simulates the light rays and object reflectance, and finally renders the haze effects, which demonstrates their superiority over state-of-the-art methods in terms of both image quality and runtime.
- Learning Affordance Grounding from Exocentric Images142
A cross-view knowledge transfer framework that extracts affordance-specific features from exocentric interactions and enhances the perception of affordance regions by preserving affordance correlation is devised.
- 142
- Step-Audio: Unified Understanding and Generation in Intelligent Speech Interaction129
Step-Audio is introduced, the first production-ready open-source solution that achieves state-of-the-art performance in human evaluations, especially in terms of instruction following and on open-source benchmarks like LLaMA Question.
- Progressive Retinex125
A novel progressive Retinex framework is presented, in which illumination and noise of low-light image are perceived in a mutually reinforced manner, leading to noise reduction low- light enhancement results.
- 117
- Automatic Pavement-Crack Detection and Segmentation Based on Steerable Matched Filtering and an Active Contour Model108
This data indicates that pavement cracks are an important symptom of pavement deterioration and deficiency and accurate and complete information regarding pavement cracks is critical to determining pavement maintenance schedul...
- Grounding 3D Object Affordance from 2D Interactions in Images90
A novel Interaction-driven 3D Affordance Grounding Network (IAG), which aligns the region feature of objects from different sources and models the interactive contexts for 3D object affordance grounding and collects a Point-Image Affordance Dataset (PIAD) to support the proposed task.
- 85
- 79
- 3D-QSAR Model of Flavonoids Binding at Benzodiazepine Site in GABAA Receptors79
With flavone as a structural template, three-dimensional quantitative structure-activity relationship (3D-QSAR) studies and ab initio calculations were performed on a series of flavonoids to build a reasonable pharmacophore model and propose specific S1 and S2 subsites relating to substituent positions.
- 73
- 69
- 68
- E-MLB: Multilevel Benchmark for Event-Based Camera Denoising65
A large-scale event denoising dataset is constructed for the first time and the first nonreference event Denoising metric, the event structural ratio (ESR), which measures the structural intensity of given events is proposed, which is inspired by the contrast metric.
- LEMON: Learning 3D Human-Object Interaction Relation from 2D Images61
LEMON (LEarning 3D huMan-Object iNteraction relation), a unified model that mines interaction intentions of the counter-parts and employs curvatures to guide the extraction of ge-ometric correlations, combining them to anticipate the interaction elements.
- Lightweight Adaptive Feature De-Drifting for Compressed Image Classification61
A novel lightweight adaptive feature de-drifting module (AFD-Module) to boost the performance of pre-trained image classification models when facing compressed images and comprehensively improve the accuracy of the pre-trained classification models and significantly outperform the existing methods.
- One-Shot Object Affordance Detection in the Wild61
A One-Shot Affordance Detection Network (OSAD-Net) is devised that firstly estimates the human action purpose and then transfers it to help detect the common affordance from all candidate images and builds a large-scale purpose-driven affordance dataset v2.
- 61
- 59
- 58
- 58
- DisasterGAN: Generative Adversarial Networks for Remote Sensing Disaster Image Generation58
Two models are proposed: disaster translation GAN can generate disaster images for multiple disaster types using only a single model, which uses an attribute to represent disaster types and a reconstruction process to further ensure the effect of the generator; damaged building generation GAN is a mask-guided image generation model,Which shows the effectiveness of the proposed models and the superiority compared with other data augmentation methods.
- 58
- Deep Learning Neural Network Model for Tunnel Ground Surface Settlement Prediction Based on Sensor Data53
Experimental results show that, compared with existing machine learning techniques and algorithms, this deep learning method has higher prediction accuracy and acceptable computational efficiency, and in the case of small samples, this method can significantly improve the accuracy of time series forecasting.
- Background Activation Suppression for Weakly Supervised Object Localization and Semantic Segmentation48
A background activation suppression (BAS) method designed to facilitate the learning of generator by suppressing the background activation value and which achieves state-of-the-art weakly supervised semantic segmentation performance on the PASCAL VOC 2012 and MS COCO 2014 datasets.
- 47
- Object affordance detection with relationship-aware network47
A novel relationship-aware convolutional neural network, which takes the symbiotic relationship between multiple affordances and the combinational relationship between the affordance and objectness into consideration, to predict the most probable affordance label for each pixel in the object.
- Learning Visual Affordance Grounding From Demonstration Videos45
A hand-aided affordance grounding network that leverages the aided clues provided by the position and action of the hand in demonstration videos to eliminate the multiple possibilities and better locate the interaction regions in the object.
- 45
- 45
- 43
- Deep time–frequency representation and progressive decision fusion for ECG classification43
A novel ECG classification method is proposed by learning deep time-frequency representation and progressive decision fusion at different temporal scales in an end-to-end manner to fuse decisions from the scale-specific models into a more accurate and stable one.
- Deep spatiotemporal residual early-late fusion network for city region vehicle emission pollution prediction41
It is demonstrated that combining deep spatiotemporal residual early-late fusion network with the semi-supervised geographical weighted regression can predict vehicle emission in each region of city effectively.
- One-Shot Affordance Detection40
A One-Shot Affordance Detection (OS-AD) network that firstly estimates the purpose and then transfers it to help detect the common affordance from all candidate images to empower robots with this ability in unseen scenarios.
- 40
- Efficient Real-World Image Super-Resolution Via Adaptive Directional Gradient Convolution38
A novel equivalent parameter fusion method for DGConv is devised that maintains its rich representational capabilities while keeping computational costs consistent with a single Vanilla Convolution (VConv), enabling DGConv to improve the performance of existing super-resolution networks without incurring additional computational expenses.
- Fully Point-wise Convolutional Neural Network for Modeling Statistical Regularities in Natural Images38
This paper proposes to randomly shuffle the pixels in the origin images and leverage the shuffled image as input to make CNN more concerned with the statistical properties and shows the superiority of the proposed network over the existing architectures, i.e., using 1/10~1/100 network parameters and computational cost while achieving comparable performance.
- Leveraging Deep Statistics for Underwater Image Enhancement37
A convolutional neural network (CNN)-based framework to learn hierarchical statistical features related to color cast and contrast degradation and to leverage them for underwater image enhancement is proposed and achieves state-of-the-art performance.
- Boosting Image De-Raining via Central-Surrounding Synergistic Convolution36
Evaluations of twelve de-raining methods on nine public datasets demonstrate that the proposed Synergistic Convolution can comprehensively improve the performance of twelve de-raining networks under various rainy conditions without changing the original network structure or introducing extra computational costs.
- Effective Single Underwater Image Enhancement by Fusion36
Since the algorithm does not employ deconvolution (computationally expensive), the algorithm reduce the execution time and can effectively enhance the underwater image.
- Brightness Perceiving for Recursive Low-Light Image Enhancement34
The decompose low-light image enhancement (LLIE) into a recursive enhancement task and propose a brightness perceiving-based recursive enhancement framework for high dynamic range LLIE, which achieves new state-of-the-art (SOTA) performance on six reference and no-reference metrics.
- Grounded Affordance from Exocentric View33
A cross-view affordance knowledge transfer framework that extracts affordance-specific features from exocentric interactions and transfers them to the egocentric view to solve the above problems and the perception of affordance regions is enhanced by preserving affordance co-relations.
- 33
- 31
- Phrase-Based Affordance Detection via Cyclic Bilateral Interaction31
To perceive affordances from a vision-language perspective, and consider the challenging phrase-based affordance detection task, a cyclic bilateral consistency enhancement network (CBCE-Net) is proposed to align language and vision features in a progressive manner.
- Self-Supervised Tuning for Few-Shot Segmentation30
An adaptive tuning framework is presented, in which the distribution of latent features across different episodes is dynamically adjusted based on a self-segmentation scheme, augmenting category-specific descriptors for label prediction.
- 29
- 29
- Automatic chessboard corner detection method29
An automatic chessboard corner detection algorithm is presented for camera calibration that only requires a user input of the chessboard size, while all the other parameters can be adaptively calculated with a statistical approach.
- Categorical matrix completion28
Borders for categorical matrix completion are optimal up to a factor on the order of the square root of the number of categories, which is consistent with an intuition that the problem becomes harder when the numberof categories increases.
- 27
- PixTextGAN: structure aware text image synthesis for license plate recognition27
Qualitative and quantitative experiments demonstrate the superiority of authors’ proposed method in text image synthetisation over state-of-the-art GANs and demonstrate that using the synthesised images by PixTextGAN can greatly improve the recognition accuracy.
- 27
- VMAD: Visual-Enhanced Multimodal Large Language Model for Zero-Shot Anomaly Detection26
VMAD (Visual-enhanced MLLM Anomaly Detection), a framework that enriches MLLM with visual IAD knowledge through two key components: a Defect-Sensitive Structure Learning scheme that transfers patch-similarities for improved discrimination, and a Locality-enhanced Token Compression that leverages multi-level local features for fine-grained detection.
- 26
- 25
- 25
- DILRS: Domain-Incremental Learning for Semantic Segmentation in Multi-Source Remote Sensing Data24
This model aims to incrementally learn a new domain while preserving its performance on previous domains without accessing previous domain data, and has a unique parameter learning structure that reparametrizes domain-agnostic and domain-specific parameters.
- 24
- 23
- 23
- 22
- Sketching for sequential change-point detection22
The relative performance measure of the sketching procedure compared to that without sketching is characterized and it is shown that there can be little performance loss when the signal strength is sufficiently large, and enough number of sketches are used.
- Spatial-Aware Token for Weakly Supervised Object Localization21
This work proposes to learn a task-specific spatial-aware token (SAT) to condition localization in a weakly supervised manner and achieves state-of-the-art performance on both CUB-200 and ImageNet, with 98.45% and 73.13% GT-known Loc, respectively.
- MMAR: Towards Lossless Multi-Modal Auto-Regressive Probabilistic Modeling20
A novel Multi-Modal Auto-Regressive (MMAR) probabilistic modeling framework, which disentangle the diffusion process from auto-regressive backbone model by employing a lightweight diffusion head on top each auto-regressed image patch embedding.
- 20
- An effective approach for land-cover classification from airborne lidar fused with co-registered data20
An effective approach in which lidar data fused with its co-registered images and other derived features are used effectively for accurate urban land-cover classification and the experimental results prove that fusing the height data and optical images is particularly suited for urban Landcover classification.
- 19
- 19
- Materials structure-property prediction using a self-architecting neural network18
This paper proposes the use of a self-architecting neural network to model the relation between materials structure and properties for the purpose of predicting the properties of new materials, i.e. to predict properties for an unknown compound.
- 16
- LA-Net16
A novel Layout-Aware Convolutional Neural Network (LA-Net) is proposed for accurate monocular depth estimation by simultaneously perceiving scene layout and local depth details and formulate an order-sensitive softmax loss to well constrain the ill-posed depth inferring problem.
- Enhancing Federated Domain Adaptation with Multi-Domain Prototype-Based Federated Fine-Tuning15
Empirical results show that MPFT significantly improves both in-domain and out-of-domain accuracy over conventional methods, enhancing knowledge preservation and adaptation in FDA.
- 15
- 15
- SM-CycleGAN: crop image data enhancement method based on self-attention mechanism CycleGAN14
The potential of the self-attention mechanism on crop image datasets is explored and an innovative crop image data-enhancement method for recurrent generative adversarial networks (GANs) fused with the self-attention mechanism to significantly enhance the perception and information capture capabilities of recurrent GANs is proposed.
- 14
- Location-Free Camouflage Generation Network14
A novel Location-free Camouflage Generation Network (LCG-Net) that fuse high-level features of foreground and background image, and generate result by one inference is proposed, which is hundreds of times faster than previous methods.
- PEAR: phrase-based hand-object interaction anticipation13
A novel model, PEAR (phrase-based hand-object interaction anticipation), is proposed, which jointly anticipates interaction intention and manipulation and establishes bidirectional constraints between intention and manipulation using dynamic integration and residual connections, thereby overcoming manipulation uncertainty.
- Physics-informed deep Koopman operator for Lagrangian dynamic systems13
The proposed energy-preserving deep Lagrangian Koopman (EPDLK) framework effectively represents the dynamics of the Lagrangian system while ensuring control consistency and achieves better control consistency by guaranteeing energy difference matching.
- 13
- Vehicle emission control on road with temporal traffic information using deep reinforcement learning13
A deep reinforcement learning emission control strategy, which automatically learns the optimal traffic flow and speed limits to reduce traffic emission on the target road segment based on the temporal traffic information is presented.
- 13
- A networked remote sensing system for on-road vehicle emission monitoring13
This work discusses the principle and structure of three different types of remote sensing detectors, and also the key techniques to establish a networked remote sensing monitoring system.
- 13
- 13
- 12
- 12
- 11
- 11
- Traffic emission estimation under incomplete information with spatiotemporal convolutional GAN11
A two-step emission estimation model under incomplete information, which exploits the spatiotemporal propagation features of emission information, and generates a more reasonable spatial and temporal distribution of vehicle emissions than state-of-the-art methods.
- 11
- Prediction of Spatiotemporal Evolution of Urban Traffic Emissions Based on Taxi Trajectories11
A joint framework to predict the future evolution of vehicle emissions based on the GPS trajectories of taxis with a multi-channel spatiotemporal network and an MOVES model to estimate vehicle emissions by integrating several traffic factors including the predicted traffic states, road networks and the statistical information of urban vehicles.
- Man-machine verification of mouse trajectory based on the random forest model11
This study proposes a random forest (RF) model for man-machine verification based on the mouse movement trajectory dataset and compares it with baseline models based on performance metrics such as precision, recall, false positive rates, false negative rates, F-measure, and weighted accuracy.
- Deep Palette-Based Color Decomposition for Image Recoloring with Aesthetic Suggestion11
A deep color decomposition based framework for image recoloring is presented, allowing users to achieve professional color edition through simple interactive operations, and a novel aesthetic constraint derived from color harmony theory is proposed to augment the color reconstruction from user-specified colors.
- Theoretical studies on opioid receptors and ligands. I. Molecular modeling and QSAR studies on the interaction mechanism of fentanyl analogs binding to ?-opioid receptor11
Based on the three-dimensional models, the possible interaction mechanism of fentanyl analogs with μ-opioid receptor can be illustrated and the available structure–activity relationship of these analgesic agents can be explained reasonably.
- A tri‐attention enhanced graph convolutional network for skeleton‐based action recognition10
Experimental results demonstrate that the proposed tri ‐ attention module (TAM) can be easily plugged into existing GCNs and achieve comparable performance with the state of the art methods.
- 10
- 9
- MTTF: a multimodal transformer for temperature forecasting9
A multimodal transformer model for temperature forecasting that not only uses observational data (time series data) as input but also incorporates spatially rich ERA5 reanalysis temperature data (spatio-temporal data) and has two distinctive features.
- Effective enhancement method of low-light-level images based on the guided filter and multi-scale fusion9
An effective LLL image enhancement method based on the guided filter and multi-scale fusion for contrast enhancement and detail preservation and a new method to generate a sequence of artificial images to adjust the brightness of the output has a better performance in image detail preservation compared with other single-input algorithms.
- Deep CNN Identifier for Dynamic Modelling of Unmanned Helicopter9
Inspired by the feature expression ability of deep learning, Deep convolutional neural networks (CNN) are used to represent dynamics model of helicopter to achieve strong robustness and high accuracy, even for untrained aerobatic maneuvers.
- A novel segmentation based video-denoising method with noise level estimation9
A fast bilateral-median-filter-based method is proposed herein, which creates a segmentation-based, customised BM3D method that can be used to remove coloured multiplicative noise from videos.
- Bilateral Markov mesh random field and its application to image restoration9
Bilateral Markov mesh random field is introduced to overcome the shortcomings of the conventional Markov random fields in image modeling and is used to derive an image model to address the above shortcomings.
- 9
- High-emitter identification for heavy-duty vehicles by temporal optimization LSTM and an adaptive dynamic threshold8
This paper proposes a temporal optimization long short-term memory (LSTM) and adaptive dynamic threshold approach to identify heavy-duty high-emitters by using OBD data, which can continuously track and record the emission status in real time.
- Deep learning segmentation of the tear fluid reservoir under the sclera lens in optical coherence tomography images8
Investigating whether deep learning can be used to segment the TFR from healthy and keratoconus eyes, with irregular corneal surfaces, in OCT images found FMFE-Unet was found to outperform the other two state-of-the-art methods and ablation models, suggesting its strength in segmenting the T FR under the sclera lens depicted on OCT images.
- 8
- 8
- A Higher-Order Graph Convolutional Network for Location Recommendation of an Air-Quality-Monitoring Station8
An effective air-quality inference model-based proposed high-order graph convolution (HGCNInf) that could capture the spatiotemporal interaction of air- quality distribution and could extract external influential factor features is designed.
- 8
- Emission stations location selection based on conditional measurement GAN data8
A data augmentation strategy based on Generative Adversarial Networks (GAN), which leverages prior model COPERT with conditional measurement data to filling the missing entries at other locations to generate realistic emission data while accelerating the training process.
- 8
- Co-occurrent Structural Edge Detection for Color-Guided Depth Map Super-Resolution8
An edge detection convolutional neural network is designed to obtain the co-occurrent structural edge in depth map and its corresponding color image and is packed into another customized CNN for depth map super-resolution.
- Research on Image Acquisition and Recognition for Stored Grain Pests8
An image dataset of 9 species of pests was built up by finding the MSERs through a trap of stored grain pests combined with a real-time imaging device, and the localization and recognition of storedgrain pests were achieved.
- 8
- 8
- A Close-Form Iterative Algorithm for Depth Inferring from a Single Image8
A closeform iterative algorithm to process multi-scale image segmentation and depth inferring alternately, which can significantly improve segmentations and depth estimate results and extends the method to estimate the depth of the scene with fore-objects.
- 7
- 7
- Hypercorrelation Evolution for Video Class-Incremental Learning7
A hierarchical aggregation strategy is proposed, in which hierarchical matching matrices are combined and jointly optimized to selectively store and retrieve relevant features from previous tasks to enhance the adaptability on evolved actions.
- 7
- 7
- 7
- 7
- 7
- Sparse representation-based detail-injection method for pan-sharpening7
This work proposes a sparse representation-based detail-injection model for pan-sharpening that utilizes the structure similarity and detail differences between PAN and low-resolution multispectral (LRM) images at each channel, to improve the performance of pan- sharpening.
- 7
- Adaptively Multi-Objective Adversarial Training for Dialogue Generation7
A novel adversarial dialogue generation framework with multiple discriminators that excel in different objectives for multiple linguistic aspects, called AMPGAN, whose feasibility is proved by theoretical derivations and an adaptively adjusted sampling distribution is designed.
- One-Shot Texture Retrieval with Global Context Metric7
This paper presents an OS-TR network to encode both reference and query image, leading to texture segmentation towards the reference category, and proposes a directionality-aware module to capture the texture variations at each direction, resulting in spatially invariant representation.
- 7
- 7
- Salient object detection and classification for stereoscopic images7
An iterative method that can mutually boost salient object detection and object classification is proposed for stereoscopic images and the experimental results demonstrate the effectiveness of the method.
- Effective method for low-light image enhancement based on the JND and OCTM models6
An effective method based on the just-noticeable-difference (JND) and the optimal contrast-tone mapping (OCTM) models for low-light image enhancement that outperforms state-of-the-art methods qualitatively and quantitatively.
- 6
- 6
- Sequential Adaptive Detection for In-Situ Transmission Electron Microscopy (TEM)6
Adapt CUSUM and adaptive SSRS procedures are developed, which are based on likelihood ratio statistics with post-change mean vector being online maximum likelihood estimators with $\ell_{1}$.
- ScLNet: A cornea with scleral lens OCT layers segmentation dataset and new multi-task model5
The proposed ScLNet achieves high accuracy in extracting ScL, TFR, and corneal layer masks and boundaries from OCT images of the dataset, as evidenced by high IoU scores.
- 5
- 5
- 5
- 5
- Crowd Distribution Estimation with Multi-scale Recursive Convolutional Neural Network5
A multi-scale recursive convolutional neural network (MRCNN) based framework to map the image to its crowd distribution map and formulates two joint objectives, the estimation of crowd density map and perspective map, which enables learning a strong feature representation for both tasks.
- 5
- A new closed loop method of super-resolution for multi-view images5
A closed loop method to resolve the multi-view super-resolution problem, where the input is one high-resolution view along with its neighboring low-resolution views, that can give the super- resolution results and obtain a high-quality depth map simultaneously.
- RobustVisH: Robust Visual-Haptic Cross-Modal Recognition under Transmission Interference4
A Robust Visual-Haptic recognition (RobustVisH) model that identifies Low-Quality visual-haptic data with transmission distortion for the first time and employs a covariate shift paradigm, transferring knowledge of RobustVisH from HQ data to LQ data, thereby increasing its robustness against transmission-interference inputs.
- 4
- A steganographic backdoor attack scheme on encrypted traffic4
A steganographic backdoor attack approach based on the Least Significant Bit (LSB) to attack encrypted traffic is extended and indicates that it has an effect on the stealthiness and effectiveness of the attack.
- A seq2seq learning method for microscopic emission estimation of on-road vehicles4
This paper proposes a peak-sensitive microscopic emission estimation framework characterized by sequence-to-sequence learning for on-road vehicles and aims at capturing context features related to high emissions from sequences dynamically in an adaptive and self-learning way.
- 4
- Deep representation-based packetized predictive compensation for networked nonlinear systems4
A deep representation-based predictive compensation method is proposed for networked nonlinear systems with packet-dropouts in the feedback and forward channel based on feedback compensation and does not require a nominal system model, so that it can avoid the coupling of system uncertainties and network effects.
- 4
- Prediction of SINS/GPS Navigation Information by ELM Algorithm during GPS outages4
This paper proposes ELM for the first time to predict the navigation information on the missile-borne platform for its simpler neural network structure and no iterative calculation, and reveals that the ELM algorithm is more time-saving with higher prediction accuracy than the others.
- A Novel 3D U-Net Deep Network with Paralleling Structure for Stroke Lesion Image Segmentation4
A novel paralleling structure based on conventional 3D U-net deep network for improving the performance of CT image segmentation is proposed and the result verified that paralleling structure was useful for the convergence of loss curve.
- 4
- Depth recovery from a single defocused image based on depth locally consistency4
This paper's algorithm is based on depth local consistency assumption that the scene in input image can be modeled as multiple planar surfaces, and each over-segmented patch corresponds to a small p lanar surface in the 3D scene.
- Effective Improvement for Depth Estimated Based on Defocus Images4
Experimental results on a variety of images show that the method can acquire visual pleasing results and can also extract the depth map of the scene with fairly good extent of accuracy.
- Artificial intelligence-based segmentation of small renal masses: a multi-center, multi-scanner, multi-sequence study3
The implementation of automated segmentation achieved encouraging outcomes in both correct-detection rates and segmentation accuracy across a diverse cohort spanning multiple centers and scanners, suggesting its potential as a key component of future diagnostic pipelines for SRMs.
- 3
- 3
- Transfer learning with a spatiotemporal graph convolution network for city flow prediction3
This work proposes a transfer learning method based on spatiotemporal graph convolution, in which a dynamic spatiotemporal graph convolution module along with a temporal encoder is devised to simultaneously capture the concurrent spatiotemporal features, which implies the inherent relationship among the road network structures, human travel habits, and city bike flow.
- 3
- 3
- 3
- 3
- 3
- WDBSTF: A Weighted Dual-Branch Spatiotemporal Fusion Network Based on Complementarity between Super-Resolution and Change Prediction3
A weighted dual-branch spatiotemporal fusion network based on complementarity between super-resolution and change prediction (WDBSTF), which includes the SR branch and CP branch, and a weight module representing the complementarity of the two branches.
- 3
- 3
- 3
- 3D Layout encoding network for spatial‐aware 3D saliency modelling3
The authors first generate coarse depth-induced saliency cues which are careless of depth details, then embed both low-level and high-level RGB deep features to facilitate the spatial-aware saliency prediction.
- [An improved classification approach based on spatial and spectral features for hyperspectral data].3
An improved classification approach based on Markov random field (MRF) theory is proposed in this work and the results illuminate that the average classification accuracy of the method reachs to 95.78%, much higher than that of the result by the traditional MRF classification algorithms.
- A Virtual Network Embedding Algorithm Based on Graph Theory3
A new virtual network embedding algorithm based on node similarity, which means the similarity between the virtual nodes and the substrate nodes, achieves better mapping results between virtual network and the substr network on the topology aspect.
- Entropy analysis for inter-domain Internet application flows3
Based on detailed flow data collected from public network of Internet service provider, the statistical characteristics of applications are investigated from complex network perspective and heterogeneity of these flow graphs implies the heterogeneous behavior of users.
- 3
- Analysis of overlay topology of peer-to-peer applications3
Overlay topology properties of different P2P applications from complex network perspective are studied, including BT, Thunder, and UUSee, and degree distribution, reachability and PageRank distribution are analyzed.
- 3
- 3
- Staggered Batch Scheduling: Co-optimizing Time-to-First-Token and Throughput for High-Efficiency LLM Inference2
Staggered Batch Scheduling (SBS) is proposed, a mechanism that deliberately buffers requests to form optimal execution batches to eliminate internal queuing bubbles without compromising throughput and a Load-Aware Global Allocation strategy that balances computational load across DP units for both Prefill and Decode phases.
- Raw event-based adversarial attacks for Spiking Neural Networks with configurable latencies2
This paper proposes a novel adversarial attack method on raw event streams with configurable latencies that consistently outperforms existing approaches, achieving higher attack success rates (ASR) across various latencies.
- TLCE: Transfer-Learning Based Classifier Ensembles for Few-Shot Class-Incremental Learning2
This work uses episodic training to map images from old classes to quasi-orthogonal prototypes, which minimizes interference between old and new classes, and incorporates the use of ensembling diverse pre-trained models to further tackle the challenge of data imbalance and enhance adaptation to novel classes.
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- A Multilevel Trusted Clustering Mechanism for the Awareness Layer of the Internet of Things2
The proposed mechanism combines multi-level clustering, identity authentication, and trusted computing to solve network expansion problems, reduce energy consumption when ordinary nodes transmit data to aggregation nodes, and ensure that the identity of IOT nodes and the messages are trusted.
- 2
- Dynamic Focus Capture Method and Its Application2
The main task of the motion capture is to analyze and process the data from the sensor, and abstract to recognize the action.
- 2
- 2D/3D Model-Based Facial Video Coding/Decoding at Ultra-Low Bit-Rate2
At ultra-low bit-rate, the object experiment confirmed the advantage between coding efficiency and decoding quality of it, and the between subjects experiment indicated the suitability of subjective face identification by it.
- 2
- A Novel Hierarchical Model of Attention: Maximizing Information Acquisition2
A novel attention model to produce saliency maps by generating information distributions on incoming images by automatically marks spots with large information amount as saliency, which ensures the system gains the maximum information acquisition through attending these spots.
- ANN-Based Multi-scales Prediction of Self-similar Network Traffic2
The result shows that this algorithm can predict the self-similar network traffic at multi-scale, which is very useful to optimize network control scheme and improve network performance.
- 2
- DPSO-Based Clustering Routing Approach for WSN2
The results show that DPSOCA outperforms LEACH with significantly prolonging the networks lifetime and efficiently balancing the networks energy-dissipation, and that a greater randomicity of inertia weight and k-convergence criterion achieve a higher cost of performance of network lifetime to convergence iterations.
- 1
- 1
- 1
- 1
- 1
- 1
- A review of deep learning methods for mobile source emissions prediction1
This paper provides a comprehensive review of recent work on mobile source emissions prediction using deep learning methods and provides a deeper discussion of the future prospects and ongoing challenges.
- Two-stream spatiotemporal image fusion network based on difference transformation1
This work proposes a two-stream difference transformation spatiotemporal fusion (TSDTSF), which includes transformation and fusion streams, and shows superior performance in accuracy, vision quality, and robustness.
- 1
- 1
- 1
- The Deep Input-Koopman Operator for Nonlinear Systems1
A method that exploit the Koopman operator theory to make the strongly nonlinear dynamical systems approximately represented in the linear framework based on deep neural network (DNN) which is data-driven and equation-free is proposed.
- Effective Clone Detection Model Based on Compressed Sensing in Clustered WSNs1
A clone detection method based on compressed sensing in a cluster network is proposed that combines the local and global detection of nodes and shows that the detection rate and communication cost are superior to other methods.
- Packet-Dropouts Compensation for Networked Control System via Deep ReLU Neural Network1
Simulation results indicate that the proposed packet-dropouts compensation strategy can achieve much better control performances than the widely used zero-input or hold-input strategies, especially when the system inputs include abundant noises.
- 1
- 1
- Notice of Removal Blur kernel estimation using blurry structure1
This paper will first exploit a mask determined by convolution of an intermediate image with a kernel to generate blurry structure, and then take it into data term instead of blurry image to overcome the problem.
- 1
- Query-by-sketch Semantic Object Image Retrieval1
The experimental results show that the proposed query-by-sketch and relevance feedback based semantic object image retrieval system is effective in retrieving images and can mark semantic object in the image.
- –
- –
- –
- Learning Object Affordance Ranking with Task Context–
The first large-scale task-oriented affordance ranking dataset is constructed and the feasibility of the task context based affordance learning paradigm is demonstrated and the superiority of the model over state-of-the-art models in the fields of saliency ranking and multimodal object detection is demonstrated.
- Language model optimization and corpus construction techniques for Japanese speech recognition–
It shows that the language model optimization method and corpus construction technique chosen in this paper can improve the generalization ability of the Japanese speech recognition model.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- Near-optimality of sequential joint detection and estimation via online mirror descent.–
This work proves the optimality by making a connection of sequential analysis with online convex optimization and leveraging the logarithmic regret bound property of online mirror descent algorithm.
- –
- –
- Improving Color Constancy with Internet Photo Collections–
An iterative weighted Specular-Edge color constancy scheme that leverages large database of images gathered from the web and uses the visual context of an input image as an image-specific prior to increase the accuracy of estimating the illuminant.
- –
- –
- –
- –
- –
- Storage Reliability Grey Prediction Model for Missile Equipment–
Aiming to deal with the problem of storage reliability prediction for missile equipment, corresponding grey prediction model was established and a new method of estimation of the model parameters was proposed,Kalman filtering algorithm was used to reduce the randomicity of the original data.
- –
- –
- Working Group: Architecture, Signaling, OAM&P TITLE: Address Resolution Proposal–
An address resolution method to associate addresses in the client space with those in the optical network, which supports the overlay network model and uses the notion of a logical address for termination points.
- Monitoring and Control of Rugate Filter Fabrication Using the Orthogonal Functional Basis Neural Network–
The results of this theoretical study demonstrate that the OFBNN can achieve significantly higher prediction accuracies than the simple curve fitting technique currently in use.
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 2026-10-15. 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.