Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks111 from public data
- Geometric Transformer for Fast and Robust Point Cloud Registration514
This work proposes Geometric Transformer, a simplistic design that attains surprisingly high matching accuracy such that no RANSAC is required in the estimation of alignment transformation, leading to 100 times acceleration.
- GeoTransformer: Fast and Robust Point Cloud Registration With Geometric Transformer303
Geometric Transformer is proposed, or GeoTransformer for short, to learn geometric feature for robust superpoint matching, making it invariant to rigid transformation and robust in low-overlap cases.
- ThunderNet: Towards Real-Time Generic Object Detection on Mobile Devices292
benefit from the highly efficient backbone and detection part design, ThunderNet surpasses previous lightweight one-stage detectors with only 40% of the computational cost on PASCAL VOC and COCO benchmarks.
- Similarity measures between intuitionistic fuzzy (vague) sets: A comparative analysis279
The research presented here could benefit selection and applications of similarity measures for intuitionistic fuzzy sets and vague sets in practice.
- Rotation-Invariant Transformer for Point Cloud Matching179
This work introduces RoITr, a Rotation-Invariant Transformer to cope with the pose variations in the point cloud matching task, and proposes a global transformer with rotation-invariant cross-frame spatial awareness learned by the self-attention mechanism, which significantly improves the feature distinctiveness and makes the model robust with respect to the low overlap.
- MotionTrack: Learning Robust Short-Term and Long-Term Motions for Multi-Object Tracking172
A simple yet effective multi-object tracker, i.e., MotionTrack, which learns robust short-term and long-term motions in a unified framework to associate trajectories from a short to long range, is proposed.
- Enabling Efficient Cyber Threat Hunting With Cyber Threat Intelligence140
The proposed ThreatRaptor is an unsupervised, light-weight, and accurate NLP pipeline that extracts structured threat behaviors from unstructured OSCTI text and provides a concise and expressive domain-specific query language, TBQL, to hunt for malicious system activities.
- Fd-Mobilenet: Improved Mobilenet with a Fast Downsampling Strategy113
Fast-Downsampling MobileNet is presented, an efficient and accurate network for very limited computational budgets (e.g., 10–140 MFLOPs) that consistently outperforms MobileNet and achieves comparable results with ShufflieNet under different computational budgets.
- 112
- 98
- RIGA: Rotation-Invariant and Globally-Aware Descriptors for Point Cloud Registration88
RIGA is introduced to learn descriptors that are Rotation-Invariant by design and Globally-Aware to learn descriptors that are Rotation-Invariant by design and Globally-Aware for point cloud registration.
- 88
- 86
- SC2Net: Sparse LSTMs for Sparse Coding61
This work recasts ISTA as a recurrent neural network unit and shows its connection with the well-known long short term memory (LSTM) model, and presents a neural network to achieve sparse codes in an end-to-end manner.
- Towards Generalizable Multi-Object Tracking58
This paper investigates the factors that influence trackers' generalization to different scenar-ios and concretize them into a set of tracking scenario at-tributes to guide the design of more generalizable trackers and proposes a "point-wise to instance-wise relation" framework for MOT, i.e., GeneralTrack, which can generalize across diverse scenarios while eliminating the need to balance motion and appearance.
- 58
- Task-Specific Context Decoupling for Object Detection57
This work proposes a novel Task-Specific COntext DEcoupling (TSCODE) head which further disentangles the feature encoding for two tasks, and provides high-resolution feature map containing more edge information to better regress object boundaries.
- Multi-Prompts Learning with Cross-Modal Alignment for Attribute-based Person Re-Identification53
A new framework called Multi-Prompts ReID (MP-ReID), based on prompt learning and language models, to fully dip fine attributes to assist ReID task with off-the-shelf models for more accurate retrieval results.
- 52
- 2D3D-MATR: 2D-3D Matching Transformer for Detection-free Registration between Images and Point Clouds43
2D3D-MATR, a detection-free method for accurate and robust registration between images and point clouds that adopts a coarse-to-fine pipeline where it first computes coarse correspondences between downsampled patches of the input image and the point cloud and then extends them to form dense correspondences between pixels and points within the patch region.
- 43
- 43
- Real-time texture-mapped vector glyphs42
A vector graphics representation suitable for real-time rendering on GPUs that supports inexpensive high-quality anisotropic antialiasing as well as special effects such as outlining (with both rounded and sharp miters) and embossing.
- Deep Graph-based Spatial Consistency for Robust Non-rigid Point Cloud Registration41
Graph-based Spatial Consistency Network (GraphSCNet) is proposed to filter outliers for non-rigid registration and effectively improves the quality of the putative correspondences and attains state-of-the-art performance on three challenging benchmarks.
- 40
- 40
- Precise vector textures for real-time 3D rendering39
A simple algorithm is presented that can efficiently and accurately compute the minimum distance to a parametric curve when the sample point is within its radius of curvature and the curve can be segmented into monotonic regions.
- 39
- 35
- Loss Rank Mining: A General Hard Example Mining Method for Real-time Detectors32
LRM is the first hard example mining strategy which could fit YOLOv2 perfectly and make it better applied in series of real scenarios where both real-time rates and accurate detection are strongly demanded.
- Diagonalwise Refactorization: An Efficient Training Method for Depthwise Convolutions31
The key idea is to rearrange the weight vectors of a depthwise convolution into a large diagonal weight matrix so as to convert the depthwise Convolution into one single standard convolution, which is well supported by the cuDNN library that is highly-optimized for GPU computations.
- Unsupervised Voice-Face Representation Learning by Cross-Modal Prototype Contrast30
The proposed cross-modal prototype contrastive learning (CMPC) takes advantage of contrastive methods and resists adverse effects of false negatives and deviate positives, and outperforms state-of-the-art unsupervised methods on various voice-face association evaluation protocols.
- Learning accurate template matching with differentiable coarse-to-fine correspondence refinement29
This work proposes an accurate template matching method based on differentiable coarse-to-fine correspondence refinement that is significantly better than state-of-the-art methods and baselines, providing good generalization ability and visually plausible results even on unseen real data.
- 27
- Delocate: Detection and Localization for Deepfake Videos with Randomly-Located Tampered Traces22
Delocate is a novel Deepfake detection model that can both recognize and localize unknown domain Deepfake videos and strategically emphasizes the recovery phase of fake faces with poor recovery, facilitating the localization of tampered regions.
- Improving TextRank Algorithm for Automatic Keyword Extraction with Tolerance Rough Set20
This paper improves TM in two aspects with TRS theory that is used to mine vocabulary, semantics, grammar and other information in the corpus, and applies the proposed methods to the clustering tasks of two datasets, outperforming the strong baselines.
- The dynamic features of Delicious, Flickr, and YouTube19
A Tagger Tag Resource Latent Dirichlet Allocation model (TTRLDA) is proposed to explore the evolution of topics emerging from social tagging vocabularies in Delicious, Flickr, and YouTube from 2003 to 2008.
- 18
- 16
- Sensor management of LEO constellation using modified binary particle swarm optimization13
A modified binary particle swarm optimization (MBPSO) algorithm is presented for use in finding the optimal allocation scheme of sensors for maximizing the tracking accuracy of the midcourse ballistic missile.
- Mover: Mask and Recovery based Facial Part Consistency Aware Method for Deepfake Video Detection11
Mover is a new Deepfake detection model that exploits unspecific facial part inconsistencies, which are inevitable weaknesses of Deepfake videos and is highly effective on standard benchmarks.
- 10
- 10
- SS-YOLOv8: A Lightweight Algorithm for Surface Litter Detection9
A lightweight network model for detecting water surface litter and enhancing the CSP Bottleneck with a two-convolutions (C2f) module to improve image recognition tasks and offers a scalable solution for deploying advanced detection models on resource-constrained devices.
- 9
- 9
- A Graphical XQuery Language Using Nested Windows9
A graphical XQuery-based language using nested windows, GXQL, is presented, which provides a powerful and user-friendly environment for non-technical users to perform queries.
- 8
- 8
- 8
- Merging and Evolution: Improving Convolutional Neural Networks for Mobile Applications8
This work presents two novel operations named merging and evolution to leverage the inter-group information and proposes a family of compact neural networks called MENet based on ME modules, which consistently outperforms other state-of-the-art compact networks under different computational budgets.
- 8
- 7
- Medical image classification algorithm based on principal component feature dimensionality reduction6
Forgery detection experiment in many image data bases indicates: WLD of multi-resolution has better detection result compared with single resolution, and WLD-of-multi-resolution have better detection performance in detection of splicing forgery images and copying–moving forgeries images.
- 6
- FPBN: A New Formalism for Evaluating Hybrid Bayesian Networks Using Fuzzy Sets and Partial Least-Squares6
This paper proposes a general formalism for evaluating hybrid Bayesian networks, called fuzzy partial least-squares Bayesian network (FPBN), which replaces each continuous variable whose descendants include discrete variables by a partner discrete variable and adding a directed link from that partner discrete variables to the continuous one.
- 5
- 5
- 5
- Merging-and-Evolution Networks for Mobile Vision Applications5
This work presents two novel operations named merging and evolution to leverage the inter-group information and proposes a family of compact neural networks called MENet based on the ME modules, which consistently outperforms other state-of-the-art compact networks under different computational budgets.
- 5
- 4
- 4
- 4
- 4
- Conversion Algorithm of Linear-Time Temporal Logic to Buchi Automata4
An improved Tableau-based algorithm, which converts LTL formulae to Transition-based Buchi Automata(TBA) more efficiently, and can expand the state nodes while detecting the validity of nodes, removing the invalid nodes and combining equivalent states and transitions.
- 4
- Recap: Detecting Deepfake Video with Unpredictable Tampered Traces via Recovering Faces and Mapping Recovered Faces3
Recover, a novel Deepfake detection model that exposes unspecific facial part inconsistencies by recovering faces and enlarges the differences between real and fake by mapping recovered faces is proposed.
- 3
- Quality prediction modelling for software customisation in the absence of defect data3
A quality prediction model which focuses on software quality prediction when defect data is unavailable is proposed and the improved model performs better than the original one in prediction accuracy and effectiveness.
- 3
- Density-aware and Cluster-based Federated Anomaly Detection on Data Streams2
This work designs a novel lightweight federated anomaly detection clusters with density-aware hash cells, which successfully capture evolving data distribution and proposes a straightforward but effective metric divergences accompanied by a greedy search algorithm, which takes both global aggregation bias mitigation and efficiency into account.
- Real‐time detection of Angelica dahurica tablet using YOLOX_am2
YOLOX_am is a novel deep learning‐based network that combines the fast detection ability of YOLOX and the feature weighting ability of the attention mechanism that is both fast and accurate for ADT's detection and can be deployed in the sorting system to meet actual production needs.
- An Experimental Perspective for Computation-Efficient Neural Networks Training2
These models have smaller CMR than the large networks, such as VGG, ResNet, Inception, etc, and can run efficiently on ARM-based devices.
- Vector Graphics for Real-time 3D Rendering2
Algorithms are presented that enable the use of vector graphics representations of images in texture maps for 3D real time rendering and an iterative algorithm is discussed that can compute the distance to any parametrically differentiable curve (including polynomial splines of any order) robustly.
- 2
- 2
- 2
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- LTL Formulae to Buchi Automata Translation Using on‐the‐Fly De‐generalization1
A conversion algorithm to translate a Linear temporal logic (LTL) formulato a Buchi automaton (BA) directly is presented and it is shown that it is more efficient for a series of LTL formulae in commonuse.
- 1
- 1
- 1
- 1
- –
- –
- –
- –
- –
- –
- –
- –
- FaceIDP: Face-IDentification Privacy under theCompressed Sensing Framework–
The Face-IDentification Privacy (FaceIDP) approach to protect the facial images from being used by the adversary to breach some FaceID system is developed and it is proved that the FaceIDP is $\varepsilon$-differentially private.
- –
- –
- The Improved Velocity-based Models for Pedestrian Dynamics–
A novel model of pedestrian dynamics is proposed using a dynamic ellipse, which is parameterized by their velocity and can improve the simulaton accuracy and found a representation of the pedestrian much closer to the reality.
- –
- Degeneralization Algorithm for Generation of Büchi Automata Based on Contented Situation–
On-the-fly degeneralization algorithm used to transform generalized Buchi automata (GBA) into Buchi Automata (BA) different from the standard degeneralizations is presented and it is shown that it is more efficient for four sets of random formulae generated by LBTT.
- –
- –
- PageRank algorithm with semantic relevance of anchor texts–
The experimental results show that the new hyperlink algorithm put forward based on the relevance of hyperlink anchor texts is effective, and it can raise the precision of PageRank algorithm.
- PROLOG for the Implementation of Student Module in ICAIP system.–
The catalytic hydrodesulfurization and/or hydrodenitrogenation of an organic sulfur compound and/ or an organic nitrogen compound is carried out in the presence of a catalyst composition comprising catalytic grade alumina, titanium, cobalt, zinc and molybdenum.
- Observational Learning Algorithm for Heterogeneous Ensembles–
Experiments on five datasets from the UCI repository show that, OLA outperforms the individual base learner and majority voting when base learners are not capable enough for the given task.
- –
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.