Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks166 from public data
- DenseCLIP: Language-Guided Dense Prediction with Context-Aware Prompting886
This work presents a new framework for dense prediction by implicitly and explicitly leveraging the pre-trained knowledge from CLIP and converts the original image-text matching problem in CLIP to a pixel- text matching problem and uses the pixel-text score maps to guide the learning of dense prediction models.
- Stochastic Trajectory Prediction via Motion Indeterminacy Diffusion438
This paper presents a new framework to formulate the trajectory prediction task as a reverse process of motion indeterminacy diffusion (MID), in which the history behavior information and the social interactions are encoded as a state embedding and a Transformer-based diffusion model is devised to capture the temporal dependencies of trajectories.
- 419
- Counterfactual Attention Learning for Fine-Grained Visual Categorization and Re-identification380
This paper analyzes the effect of the learned visual attention on network prediction through counterfactual intervention and maximize the effect to encourage the network to learn more useful attention for fine-grained image recognition.
- FineDiving: A Fine-grained Dataset for Procedure-aware Action Quality Assessment203
This work proposes a procedure-aware approach for action quality assessment, learned by a new Temporal Segmentation Attention module, and proposes to parse pairwise query and exemplar action instances into consecutive steps with diverse semantic and temporal correspondences.
- 200
- Automatic EEG seizure detection using dual-tree complex wavelet-Fourier features182
A new Electroencephalography (EEG) seizure detection method by using the dual-tree complex wavelet (DTCWT) - Fourier features, which achieve perfect classification rates (100%) for the EEG database from the University of Bonn.
- 160
- PromptCAL: Contrastive Affinity Learning via Auxiliary Prompts for Generalized Novel Category Discovery139
This work proposes contrastive affinity learning to calibrate semantic representations based on the authors' iterative semi-supervised affinity graph generation method for semantically-enhanced supervision and proposes a discriminative prompt regularization loss to reinforce semantic discriminativeness of prompt-adapted pre-trained vision transformer for refined affinity relationships.
- Invariant Fourier-wavelet descriptor for pattern recognition135
The result shows that the Fourier-wavelet descriptor is an efficient representation which can provide for reliable recognition of complex patterns such as roadsigns, keys, aircrafts, characters, etc.
- Path planning of mobile robot with PSO-based APF and fuzzy-based DWA subject to moving obstacles106
This paper proposes a two-layer path-planning method, where an optimized artificial potential field method and an improved dynamic window approach (DWA) are used at the global and local layer, respectively, to enable the robot to plan a better path under a multi-obstacle environment while avoiding the moving obstacles effectively.
- 105
- Deep Credible Metric Learning for Unsupervised Domain Adaptation Person Re-identification100
Experimental results demonstrate that the proposed deep credible metric learning (DCM-L) method explores credible and valuable training data and improves the performance of unsupervised domain adaptation.
- Human Trajectory Prediction via Counterfactual Analysis98
This work proposes a counterfactual analysis method for human trajectory prediction to investigate the causality between the predicted trajectories and input clues and alleviate the negative effects brought by the environment bias.
- PLOT: Prompt Learning with Optimal Transport for Vision-Language Models86
This work proposes to apply optimal transport to match the vision and text modalities and first model images and the categories with visual and textual feature sets and applies a two-stage optimization strategy to learn the prompts.
- Temporal Coherence or Temporal Motion: Which Is More Critical for Video-Based Person Re-identification?83
This paper proposes a simple yet effective Adversarial Feature Augmentation (AFA) method, which highlights the temporal coherence features by introducing adversarial augmented temporal motion noise.
- Automatic Epileptic Seizure Detection in EEG Using Nonsubsampled Wavelet–Fourier Features78
A method for EEG seizure detection by decomposing EEG signals for up to six wavelet scales without downsampling and taking the magnitude of the Fourier coefficients as features for seizure detection is proposed.
- IMAGE DENOISING BASED ON WAVELET SHRINKAGE USING NEIGHBOR AND LEVEL DEPENDENCY78
Three different wavelet shrinkage methods, namely NeighShrink, NeighSure and NeighLevel, are presented, which give comparatively higher peak signal to noise ratio (PSNR), are much more efficient and have less visual artifacts compared to other methods.
- Personalized Trajectory Prediction via Distribution Discrimination71
A distribution discrimination method to predict personalized motion patterns by distinguishing the potential distributions, which can be integrated with existing multi-modal stochastic predictive models as a plug-and-play module to learn the more discriminative latent distribution.
- 69
- 61
- Denoising and dimensionality reduction of hyperspectral imagery using wavelet packets, neighbour shrinking and principal component analysis61
The proposed method that can fulfil both denoising and dimensionality reduction of hyperspectral data using wavelet packets, neighbour wavelet shrinking and PCA is confirmed.
- Simultaneous dimensionality reduction and denoising of hyperspectral imagery using bivariate wavelet shrinking and principal component analysis61
Experiments reported in this paper confirm that the proposed method outperforms PCA for endmember extraction and mineral detection using the Cuprite data cube.
- Understanding Masked Autoencoders via Hierarchical Latent Variable Models57
This work formally characterize and justify existing empirical in-sights and provide theoretical guarantees of MAE, and formulate the underlying data-generating process as a hierarchical latent variable model, and shows that under reasonable assumptions, MAE provably identifies a set of latent variables in the hierarchical model.
- 56
- Dimensionality reduction of hyperspectral imagery using improved locally linear embedding55
The improved LLE can process larger hyperspectral images than the existing LLE and it is also faster and better results are obtained for detection.
- 52
- 48
- 47
- 46
- Unsupervised Sampling Promoting for Stochastic Human Trajectory Prediction44
This work proposes a novel method, called BOsampler, to adaptively mine potential paths with Bayesian optimization in an unsupervised manner, as a sequential design strategy in which new prediction is dependent on the previously drawn samples.
- 43
- Narrative Action Evaluation with Prompt-Guided Multimodal Interaction39
A prompt-guided multimodal interaction framework that outperforms separate learning methods and naive multi-task learning methods in narrative action evaluation and establishes benchmarks for NAE.
- Security improvement for asymmetric cryptosystem based on spherical wave illumination37
The proposed improvement method can avoid various types of the currently existing attacks and maintain the asymmetric characteristic of the cryptosystem due to only changing the illuminating mode and can be easily implemented in optics compared with the reported improving methods.
- Tem-adapter: Adapting Image-Text Pretraining for Video Question Answer34
The proposed Tem-adapter enables the learning of temporal dynamics and complex semantics by a visual Temporal Aligner and a textual Semantic Aligner that first designs a template to fuse question and answer pairs as event descriptions and then learns a Transformer decoder with the whole video sequence as guidance for refinement.
- Are electroencephalogram (EEG) signals pseudo-random number generators?31
This paper's binary sequences have passed nearly all NIST pseudo-random number tests except a few failures, indicating that the EEG signals can indeed become a PRNG provided that it undergoes a transformation like the authors'.
- Causality for Large Language Models29
How causality can enhance LLMs at every stage of their lifecycle is explored, from token embedding learning and foundation model training to fine-tuning, alignment, inference, and evaluation-paving the way for more interpretable, reliable, and causally-informed models.
- Hyperspectral face recognition via feature extraction and CRC‐based classifier28
Five existing descriptors that are frequently used in 2D face recognition are compared, collaborative representation classifier (CRC) is used with two voting techniques for hyperspectral face recognition, and Experimental results demonstrate that Gabor filter bank-based features are very robust to both Gaussian white noise and shot noise, and it achieves very competitive classification results.
- Preparation and characterization of PVPI-coated Fe3O4 nanoparticles as an MRI contrast agent28
Results indicated that the PVPI-coated Fe3O4 nanoparticles had great potential for application in MRI as a T2 contrast agent.
- 28
- CaRiNG: Learning Temporal Causal Representation under Non-Invertible Generation Process27
This work establishes an identifiability theory that allows for the recovery of independent latent components even when they come from a nonlinear and non-invertible mix and proposes a principled approach, CaRiNG, to learn the CAusal RepresentatIon of Non-invertible Generative temporal data with identifiability guarantees.
- Federated Causal Discovery from Heterogeneous Data25
A novel FCD method attempting to accommodate arbitrary causal models and heterogeneous data by utilizing a surrogate variable corresponding to the client index to account for the data heterogeneity across different clients and establishing a federated independent change principle to determine causal directions.
- 25
- DENOISING OF THREE-DIMENSIONAL DATA CUBE USING BIVARIATE WAVELET SHRINKING23
This paper extended Sendur and Selesnick's bivariate wavelet thresholding from two-dimensional image denoising to three-dimensional (3D) data cube denoise, and shows that bivariate waveshell thresholding is still valid for 3D data cubes.
- Partial Identifiability for Domain Adaptation20
This work proposes a practical domain adaptation framework called iMSDA, which outperforms state-of-the-art domain adaptation algorithms on benchmark datasets, demonstrating the effectiveness of this property-based framework.
- Feature Expansion for Graph Neural Networks20
This paper decomposes graph neural networks into determined feature spaces and trainable weights, providing the convenience of studying the feature space explicitly using matrix space analysis, and theoretically finds that thefeature space tends to be linearly correlated due to repeated aggregations.
- 20
- 20
- SUPER-RESOLUTION OF HYPERSPECTRAL IMAGERY USING COMPLEX RIDGELET TRANSFORM20
A novel super-resolution method for hyperspectral imagery is proposed by using complex ridgelet transform, which has the property of approximate shift invariance, which is very important in image super- resolution.
- On the Identification of Temporally Causal Representation with Instantaneous Dependence19
This work establishes identifiability results of the latent causal process based on sufficient variability and the sparse influence constraint by employing contextual information of time series data and incorporates a temporally variational inference architecture to estimate the latent variables and a gradient-based sparsity regularization to identify the latent causal process.
- Optical color image encryption using position multiplexing technique based on phase truncation operation19
This work employs the position multiplexing technique to encrypt the color image in only one spatial channel and can maintain the nonlinear characteristic of the cryptosystem and avoid various types of the currently existing attacks, especially the iterative attack.
- Denoising of hyperspectral imagery by combining PCA with block-matching 3-D filtering19
The proposed method for reducing noise in a hyperspectral data cube first used principal component analysis (PCA) to decorrelate the useful information from noise and then applied block-matching and 3-D filtering to selectively reduce noise in the noisy PCA components.
- Evaluation and comparison of dimensionality reduction methods and band selection18
The evaluation and comparison of three DR methods‐namely, principal component analysis (PCA), wavelet, and minimum noise fraction (MNF)‐and one band selection method were conducted and the following was observed.
- 16
- 15
- 14
- Hyperspectral face recognition with log‐polar Fourier features and collaborative representation based voting classifiers14
This study proposes a new method for hyperspectral face recognition that reduces noise adaptively from each spectral band by using any image denoising method, e.g. block matching and 3D filtering, and extracts 2D Fourier spectrum from them.
- MuGSI: Distilling GNNs with Multi-Granularity Structural Information for Graph Classification13
MuGSI is introduced, a novel KD framework that employs Multi-granularity Structural Information for graph classification and proposes to incorporate a node feature augmentation component, thereby enhancing the expressiveness of the student MLPs and making them more capable learners.
- A new image registration method robust to noise13
This paper estimates the translation, rotation, and scaling parameters between a reference image and a distorted image in order to register them using the ratio of means, the Radon transform, and the fast Fourier transform.
- Wavelet-based moment invariants for pattern recognition13
Experimental results show that the proposed wavelet-based moments outperform Zernike's moments and the Fourier-wavelet descriptor for pattern recognition under different rotation angles and different noise levels.
- 11
- 11
- Towards Realistic Zero-Shot Classification via Self Structural Semantic Alignment10
This paper proposes the Self Structural Semantic Alignment (S3A) framework, which extracts the structural semantic information from unlabeled data while simultaneously self-learning and proposes to self-train the CLIP image encoder with both individual and structural semantic alignment through a teacher-student learning strategy.
- 9
- 9
- Advancing Reasoning in Diffusion Language Models with Denoising Process Rewards8
This work introduces a process-level reinforcement signal defined over the denoising trajectory of diffusion language models, and proposes an efficient stochastic estimator that reuses standard training rollouts, enabling practical process-level supervision at scale.
- Reflection-Window Decoding: Text Generation with Selective Refinement8
This paper proposes an approach that incorporates a sliding reflection window and a pausing criterion, such that refinement and generation can be carried out interchangeably as the decoding proceeds, and its extensive experimental results demonstrate the effectiveness of the approach.
- 8
- 8
- 8
- 8
- 7
- 7
- 7
- 7
- Synergy Between Sufficient Changes and Sparse Mixing Procedure for Disentangled Representation Learning6
An identifiability theory with less restrictive constraints regarding distribution changes and the sparse mixing procedure is proposed, enhancing applicability to real-world scenarios and developing an estimation framework incorporating a domain encoding network and a sparse mixing constraint.
- 6
- 5
- Nonstationary Time Series Forecasting via Unknown Distribution Adaptation5
The IDEA model, which incorporates an autoregressive hidden Markov model to estimate latent environments and modular prior networks to identify latent states, out-performs several latest nonstationary forecasting methods on various benchmark datasets, high-lighting its advantages in real-world scenarios.
- 5
- 4
- Language-Free Compositional Action Generation via Decoupling Refinement4
This study introduces a novel framework to generate compositional actions without language auxiliaries using a conditional generative model, CVAE, and a self-supervised pre-trained model MAE to ensure semantic consistency between sub-actions and compositional actions.
- 4
- 4
- 4
- 4
- 4
- On a laxity-based real-time scheduling policy for fixed-priority tasks and its non-utilization bound4
- 4
- PersonaX: Multimodal Datasets with LLM-Inferred Behavior Traits3
PersonaX, a curated collection of multimodal datasets designed to enable comprehensive analysis of public traits across modalities, and introduces a novel causal representation learning (CRL) framework tailored to multimodal and multi-measurement data, providing theoretical identifiability guarantees.
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- Selection, Reflection and Self-Refinement: Revisit Reasoning Tasks via a Causal Lens2
This paper revisits reasoning tasks from a causal perspective, seeking to understand their behavior in latent space and to offer insights for addressing their challenges, and introduces a framework that incorporates the estimated latent variables as feedback into the selection mechanism, thereby facilitating the learning of dense dependencies among latent representations.
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 1
- CausalVerse: Benchmarking Causal Representation Learning with Configurable High-Fidelity Simulations1
- Learning by Analogy: A Causal Framework for Composition Generalization1
This work introduces a hierarchical data-generating process that naturally encodes different levels of concepts and their interaction mechanisms and proves that this latent hierarchical structure is provably recoverable from observable data like text-image pairs, a necessary step for learning such a generative process.
- 1
- Learning Causal Domain-Invariant Temporal Dynamics for Few-Shot Action Recognition1
DITeD is proposed, a generative framework with a two-stage training strategy during pre-training that explicitly model invariant dynamics including temporal dynamic generation and transitions, and the variant visual and domain encoders, and validate that the learned temporal dynamic transition and temporal dynamic generation modules possess transferable qualities.
- Learning Socio-Temporal Graphs for Multi-Agent Trajectory Prediction1
This work introduces a Directed Acyclic Graph-based structure, which is term Socio-Temporal Graph (STG), to explicitly capture pair-wise socio-temporal interactions among a group of people across both space and time.
- 1
- 1
- 1
- 1
- 1
- 1
- An Algorithm of Attribute Reduction Based on HU's Discernibility Matrix by Distinguishing Capability1
- 1
- 1
- 1
- 1
- –
- –
- –
- –
- –
- –
- –
- –
- MixAR: Mixture Autoregressive Image Generation–
MixAR is a novel framework that leverages mixture training paradigms to inject discrete tokens as prior guidance for continuous autoregressive prediction and proposes Training-Inference Mixture (TI-Mix) to achieve consistent training and generation distributions.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
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 2026-10-16. 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.