Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Suggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works87 from public data
- LL3DA: Visual Interactive Instruction Tuning for Omni-3D Understanding, Reasoning, and Planning290
LL3DA is presented, a Large Language 3D Assistant that takes point cloud as the direct input and responds to both text instructions and visual interactions that enables LMMs to better comprehend human interactions with the 3D environment and further remove the ambiguities within plain texts.
- $\beta$-DARTS: Beta-Decay Regularization for Differentiable Architecture Search155
A simple-but-efficient regularization method, termed as Beta-Decay, is proposed to regularize the DARTS-based NAS searching process and can impose constraints to keep the value and variance of activated architecture parameters from too large.
- DualMamba: A Lightweight Spectral–Spatial Mamba-Convolution Network for Hyperspectral Image Classification94
A novel lightweight parallel design called a lightweight dual-stream Mamba-convolution network (DualMamba) for HSI classification achieves significant classification accuracy on three public HSI datasets and a superior reduction in model parameters and floating-point operations (FLOPs).
- Landmark recognition with compact BoW histogram and ensemble ELM87
A landmark recognition framework is proposed by employing a novel discriminative feature selection method and the improved extreme learning machine (ELM) algorithm to generate a set of preliminary codewords for landmark images.
- 61
- 60
- A Large-Scale Outdoor Multi-modal Dataset and Benchmark for Novel View Synthesis and Implicit Scene Reconstruction58
A large-scale outdoor multi-modal dataset, OMMO dataset, containing complex objects and scenes with calibrated images, point clouds and prompt annotations is proposed, establishing a new benchmark for several outdoor NeRF-based tasks, such as novel view synthesis, diverse 3D representation, and multi-modal NeRF.
- 55
- 54
- Learning Cross-Image Object Semantic Relation in Transformer for Few-Shot Fine-Grained Image Classification52
Extensive experiments demonstrate that HelixFormer can effectively enhance the cross-image object semantic relation matching for recognizing fine-grained objects, achieving much better performance over most state-of-the-art methods under 1-shot and 5-shot scenarios.
- 51
- Boost Vision Transformer with GPU-Friendly Sparsity and Quantization50
A compression scheme to maximally utilize the GPU-friendly 2:4 fine-grained structured sparsity and quantization of vision transformer models that is flexible to support supervised and unsupervised learning styles.
- 49
- 45
- Performance-Aware Approximation of Global Channel Pruning for Multitask CNNs44
Experiments on several multitask datasets show that the proposed PAGCP can reduce the FLOPs and parameters by over 60% with minor performance drop, and achieves 1.3x acceleration on both cloud and mobile platforms.
- Exploring Multi-Timestep Multi-Stage Diffusion Features for Hyperspectral Image Classification41
A novel diffusion-based feature learning framework that explores multi-timestep multi-stage diffusion features for HSI classification for the first time, called MTMSD, which outperforms state-of-the-art methods for HSI classification, especially on the challenging Houston 2018 dataset.
- Object-aware Long-short-range Spatial Alignment for Few-Shot Fine-Grained Image Classification41
This work proposes an object-aware long-short-range spatial alignment approach, which is composed of a foreground object feature enhancement (FOE) module, a long-range semantic correspondence (LSC) module and a short-rangeatial manipulation (SSM) module that is developed to weaken background disturbance and encourage higher foreground object response.
- What Makes for Effective Few-shot Point Cloud Classification?34
A novel plug-and-play component called Cross-Instance Adaptation (CIA) module is proposed, to address the high intra-class variances and subtle inter-class differences issues, which can be easily inserted into current baselines with significant performance improvement.
- 32
- 29
- Point Cloud Instance Segmentation With Semi-Supervised Bounding-Box Mining28
This paper introduces the first semi-supervised point cloud instance segmentation framework (SPIB) using both labeled and unlabelled bounding boxes as supervision and introduces a novel occupancy ratio guided refinement module to refine the instance masks.
- Densely Semantic Enhancement for Domain Adaptive Region-Free Detectors28
The proposed adversarial module, namely, densely semantic enhancement module (DSEM), is proposed to strengthen the cross-domain matching of instance-level features for region-free detectors and outperforms existing domain adaptive region-based detectors under various domain shift settings.
- 28
- 24
- 24
- BridgeNet: Comprehensive and Effective Feature Interactions via Bridge Feature for Multi-Task Dense Predictions20
This work proposes a novel BridgeNet framework, which extracts comprehensive and discriminative intermediate Bridge Features, and conducts interactions based on them, and is the first work considering the completeness and quality of feature participants in cross-task interactions.
- Decouple and Orthogonalize: A Data-Free Framework for LoRA Merging19
This work proposes a Decoupled and Orthogonal merging approach (DO-Merging), which can achieve significantly higher performance than existing merging methods at a minimal cost and can be flexibly integrated with existing methods, offering near free-lunch improvements across tasks.
- Adapter-X: A Novel General Parameter-Efficient Fine-Tuning Framework for Vision17
The proposed Adapter-X is the first to outperform full fine-tuning in both 2D image and 3D point cloud modalities with significantly fewer parameters, i.e., only 0.20% and 1.88% of original trainable parameters for 2D and 3D classification tasks.
- 16
- 15
- 15
- 15
- 14
- 13
- Efficient Joint-Dimensional Search with Solution Space Regularization for Real-Time Semantic Segmentation12
A novel Solution Space Regularization (SSR) loss is first proposed to effectively encourage the supernet to converge to its discrete one and a new Hierarchical and Progressive Solution Space Shrinking method is presented to further achieve high efficiency of searching.
- EADNet: Efficient Asymmetric Dilated Network For Semantic Segmentation12
This paper proposes an efficient asymmetric dilated semantic segmentation network, named EADNet, which consists of multiple developed asymmetric convolution branches with different dilation rates to capture the variable shapes and scales information of an image.
- TBi-YOLOv5: A surface defect detection model for crane wire with Bottleneck Transformer and small target detection layer11
A surface defect detection model named TBi-YOLOv5 is proposed, inspired by the idea of bidirectional feature pyramid network (BiFPN), and a small target detection layer (STPL) is added in the neck part of YOLOV5 to detect the tiny-size surface defect on the wire rope.
- 10
- 10
- 9
- A2S-NAS: Asymmetric Spectral-Spatial Neural Architecture Search for Hyperspectral Image Classification9
This work proposes a multi-stage search architecture in order to overcome asymmetric spectral-spatial dimensions in HSI, and extends these two searchable operations to different layers of each stage to build the final architecture.
- 8
- 8
- EXIF-white balance recognition for image forensic analysis8
The proposed evidence can detect cascade manipulated images which are subject to copy-and-paste followed by different white balance post-processing operations, high compression or heavy downsampling in both theoretical and realistic scenarios.
- 8
- PDF: Point Diffusion Implicit Function for Large-scale Scene Neural Representation7
The core of the method is a large-scale point cloud super-resolution diffusion module that enhances the sparse point cloud reconstructed from several training images into a dense point cloud as an explicit prior that reduces the sampling space from the unbounded space to the scene surface.
- 7
- Coarse-to-Fine Gaze Redirection with Numerical and Pictorial Guidance6
A novel gaze redirection framework which exploits both a numerical and a pictorial direction guidance, jointly with a coarse-to-fine learning strategy and outperforms the state-of-the-art approaches in terms of both image quality and redirection precision.
- 5
- PM-INR: Prior-Rich Multi-Modal Implicit Large-Scale Scene Neural Representation5
This work explores using multi-modal priors to assist individual points to obtain more global semantic information and proposes a priorrich multi-modal implicit neural representation network, Pm-INR, for the outdoor unbounded large-scale scene.
- 5
- Spectral-spatial hyperspectral image classification using super-pixel-based spatial pyramid representation5
This work proposes a novel spectral-spatial hyperspectral image classification approach using superpixel-based spatial pyramid representation, which is a complementary yet effective feature pooling approach, and evaluation on two public hyperspectrals with superior image classification performance.
- Multipath sparse coding for scene classification in very high resolution satellite imagery5
Experimental results show that the proposed unsupervised feature learning approach for scene classification on very high resolution satellite imagery outperforms the state-of-the-art that uses the single-layer sparse coding.
- Spear: Evaluate the Adversarial Robustness of Compressed Neural Models4
The proposed Spear attack method is proposed to generate the particular adversarial attack samples for evaluating the robustness of the compressed models, and it is demonstrated the proposed Spear attack technique can generally be applied to various networks and tasks through quantitative and ablation experiments.
- Lightweight Model Pre-Training via Language Guided Knowledge Distillation4
This paper introduces language guidance to the distillation process and proposes a new method named Language-Guided Distillation (LGD) system, which uses category names of the target downstream task to help refine the knowledge transferred between the teacher and student.
- Instance-aware Model Ensemble With Distillation For Unsupervised Domain Adaptation4
A novel framework, namely Instance aware Model Ensemble With Distillation, IMED, which fuses multiple UDA component models adaptively according to different instances and distills these components into a small model is proposed.
- 4
- 4
- 3
- Rethinking Pseudo-Label-Based Unsupervised Person Re-ID with Hierarchical Prototype-based Graph3
This work proposes a novel hierarchical prototype-based graph network (HPG-Net) for unsupervised person Re-ID that outperforms state-of-the-art (SOTA) works and achieves promising performance on an occluded dataset.
- 3
- BURSTS: A bottom-up approach for robust spotting of texts in scenes3
In the proposed technique, character candidates are first detected using the proposed character detector, which leverages on the strengths of an Extremal Region detector and an Aggregate Channel Feature detector for high character detection recall.
- CTTS: Collective Test-Time Scaling2
Evaluations across seven mainstream benchmarks demonstrate that CTTS-MM significantly outperforms leading STTS methods and surpasses even flagship proprietary LLMs and open-source LLMs, highlighting the substantial potential of collective scaling to push the frontier of LLM inference.
- 2
- 2
- Attention-Aware Meta-Reweighted Optimization for Enhanced Intelligent Fault Diagnosis2
A small-sample civil aircraft fault diagnosis method that combines a meta-learning approach to tackle data imbalance with a channel attention mechanism to enhance feature extraction efficiency and demonstrates its superiority in solving imbalanced distribution experimental configurations.
- 2
- JNDMix: Jnd-Based Data Augmentation for No-Reference Image Quality Assessment2
A effective and general data augmentation based on just noticeable difference (JND) noise mixing for NR-IQA task, named JNDMix, which significantly improves the performance and data efficiency of various state-of-the-art NR- IQA models and the commonly used baseline models, as well as the generalization ability.
- MGGR: MultiModal-Guided Gaze Redirection with Coarse-to-Fine Learning2
An innovative MultiModal-Guided Gaze Redirection~(MGGR) framework that fully exploits eye-map images and target angles to adjust a given eye appearance through a designed coarse-to-fine learning that outperforms the state-of-the-art methods in terms of image quality and redirection precision.
- 2
- 2
- L-shaped corner detector for rooftop extraction from satellite/aerial imagery2
This work proposes an L-shaped corner detector for automatic rooftop extraction from high resolution satellite/aerial imagery that considers information in a spatial circle around each pixel to construct a feature map which captures the probability of L- shaped corner at every pixel.
- 1
- 1
- Spcr: semi-supervised point cloud instance segmentation with perturbation consistency regularization1
- 1
- Regions-of-interest extraction from remote sensing imageries using visual attention modelling1
The proposed visual attention model incorporates both bottom-up spatial saliency and top-down objectness, by fusing a co-occurrence histogram saliency model with the BING objectness model, and shows that the proposed model can effectively and efficiently identify regions-of-interest in remote sensing data.
- Context-aware vocabulary tree for mobile landmark recognition1
An effective mobile landmark recognition approach which exploits both context (direction and location) and content information for vocabulary tree learning and image recognition is proposed.
- 1
- 1
- –
- –
- –
- –
- –
- –
- –
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.