Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks118 from public data
- Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutions5,324
The Pyramid Vision Transformer (PVT) is introduced, which overcomes the difficulties of porting Transformer to various dense prediction tasks and is validated through extensive experiments, showing that it boosts the performance of many downstream tasks, including object detection, instance and semantic segmentation.
- PVT v2: Improved baselines with pyramid vision transformer2,673
This work improves the original Pyramid Vision Transformer (PVT v1) by adding three designs: a linear complexity attention layer, an overlapping patch embedding, and a convolutional feed-forward network to reduce the computational complexity of PVT v1 to linearity and provide significant improvements on fundamental vision tasks.
- Shape Robust Text Detection With Progressive Scale Expansion Network860
A novel Progressive Scale Expansion Network (PSENet) is proposed, which can precisely detect text instances with arbitrary shapes and is effective to split the close text instances, making it easier to use segmentation-based methods to detect arbitrary-shaped text instances.
- Stacked Conditional Generative Adversarial Networks for Jointly Learning Shadow Detection and Shadow Removal537
This paper presents a multi-task perspective, which is not embraced by any existing work, to jointly learn both detection and removal in an end-to-end fashion that aims at enjoying the mutually improved benefits from each other.
- 334
- 294
- Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection271
- YOLO-MS: Rethinking Multi-Scale Representation Learning for Real-Time Object Detection246
The core design is based on a series of investigations on how multi-branch features of the basic block and convolutions with different kernel sizes affect the detection performance of objects at different scales, and the outcome is a new strategy that can significantly enhance multi-scale feature representations of real-time object detectors.
- LSKNet: A Foundation Lightweight Backbone for Remote Sensing237
This paper considers the significance of the identified priors and proposes a lightweight Large Selective Kernel Network (LSKNet) backbone, which can dynamically adjust its large spatial receptive field to better model the ranging context of various objects in remote sensing scenarios.
- Joint Task-Recursive Learning for Semantic Segmentation and Depth Estimation237
A novel joint Task-Recursive Learning (TRL) framework for the closing-loop semantic segmentation and monocular depth estimation tasks and propagate previous learning experiences on both tasks into the next network evolution by explicitly concatenating previous responses.
- 227
- RigNet: Repetitive Image Guided Network for Depth Completion180
A repetitive design is explored in the authors' image guided network to gradually and sufficiently recover depth values and introduces a repetitive guidance module based on dynamic convolution, in which an efficient convolution factorization is proposed to simultaneously reduce its complexity and progressively model high-frequency structures.
- One-Stage Cascade Refinement Networks for Infrared Small Target Detection177
The one-stage cascade refinement network (OSCAR), which uses the high-level head as soft proposal for the low-level refinement head, allows OSCAR to process the same target in a cascade coarse-to-fine manner and demonstrates that a top-down cascade refinement framework can improve the accuracy of infrared small target detection without sacrificing efficiency.
- PAN++: Towards Efficient and Accurate End-to-End Spotting of Arbitrarily-Shaped Text177
An end-to-end text spotting framework, termed PAN++, which can efficiently detect and recognize text of arbitrary shapes in natural scenes, based on the kernel representation that reformulates a text line as a text kernel surrounded by peripheral pixels.
- Neuron-level Structured Pruning using Polarization Regularizer168
Theoretically, it is proved that polarization regularizer pushes some scaling factors to 0 and others to a value a > 0 and experimentally, it is shown that structured pruning using polarization regularizer achieves much better results than using L1 regular-izer.
- 168
- CrossKD: Cross-Head Knowledge Distillation for Object Detection160
This paper presents a general and effective prediction mimicking distillation scheme, called CrossKD, which delivers the intermediate features of the student's detection head to the teacher's detection head, and the resulting crosshead predictions are forced to mimic the teacher's predictions.
- 136
- Knowledge Distillation for Object Detection via Rank Mimicking and Prediction-Guided Feature Imitation122
Rank Mimicking (RM) and Prediction-guided Feature Imitation (PFI) are proposed for distilling one-stage detectors, respectively, which takes the rank of candidate boxes from teachers as a new form of knowledge to distill, which consistently outperforms the traditional soft label distillation.
- Regularizing Nighttime Weirdness: Efficient Self-supervised Monocular Depth Estimation in the Dark120
A novel framework with several improvements that introduces Priors-Based Regularization to learn distribution knowledge from unpaired depth maps and prevent model from being incorrectly trained and presents Statistics-Based Mask strategy to tune the number of removed pixels within textureless regions, using dynamic statistics.
- Strip R-CNN: Large Strip Convolution for Remote Sensing Object Detection95
Strip R-CNN is the first work to systematically explore large strip convolutions for remote sensing object detection, finding that strip convolutions enable directional feature aggregation along the dominant spatial dimension of slender objects, reducing background interference while preserving essential geometric information.
- DLGSANet: Lightweight Dynamic Local and Global Self-Attention Network for Image Super-Resolution83
A sparse global self-attention (SparseGSA) module is developed to select the most useful similarity values so that the most useful global features can be better utilized for image reconstruction.
- Pick of the Bunch: Detecting Infrared Small Targets Beyond Hit-Miss Trade-Offs via Selective Rank-Aware Attention76
This work proposes SeRankDet, a deep network that achieves high accuracy beyond the conventional hit-miss trade-off, by following the “Pick of the Bunch” principle, and sets new benchmarks in state-of-the-art performance across multiple public datasets.
- Densely Connected Bidirectional LSTM with Applications to Sentence Classification68
A novel multi-layer RNN model called densely connected bidirectional long short-term memory (DC-Bi-LSTM) is proposed, which essentially represents each layer by the concatenation of its hidden state and all preceding layers' hidden states, followed by recursively passing each layer's representation to all subsequent layers.
- ADNet: Lane Shape Prediction via Anchor Decomposition66
This paper revisits the limitations of anchor-based lane detection methods and introduces the Large Kernel Attention for Feature Pyramid Network (FPN) and the General Lane IoU (GLIoU) loss, which significantly improves the performance of ADNet in complex scenarios.
- SESR: Single Image Super Resolution with Recursive Squeeze and Excitation Networks60
This work proposed a novel single image super resolution method via recursive squeeze and excitation networks (SESR), which is proved to be above the state-of-the-art methods in terms of speed and accuracy.
- HazyDet: Open-Source Benchmark for Drone-View Object Detection With Depth Cues in Hazy Scenes59
Comprehensive empirical validation on HazyDet substantiates the superiority of the unified DeCoDet framework, which achieves state-of-the-art performance, surpassing the closest competitor by a notable +1.0% mAP on challenging real-world hazy test scenarios.
- Dynamic MLP for Fine-Grained Image Classification by Leveraging Geographical and Temporal Information59
This paper proposes a dynamic MLP on top of the image representation, which interacts with multimodal features at a higher and broader dimension and improves the recognizability of image representations that are visually similar but with different categories.
- Mixed Link Networks57
This work presents a highly efficient and modularized Mixed Link Network (MixNet) which is equipped with flexible inner link and outer link modules and demonstrates that MixNets can achieve superior efficiency in parameter over the state-of-the-art architectures on many competitive datasets like CIFAR-10/100, SVHN and ImageNet.
- DenoDet: Attention as Deformable Multisubspace Feature Denoising for Target Detection in SAR Images55
DenoDet is proposed, a network aided by explicit frequency-domain transform to calibrate convolutional biases and pay more attention to high frequencies, forming a natural multiscale subspace representation to detect targets from the perspective of multisubspace denoising.
- MoStGAN-V: Video Generation with Temporal Motion Styles49
It is argued that a single time-agnostic latent vector of style-based generator is insufficient to model various and temporally-consistent motions, and additional time-dependent motion styles are introduced to model diverse motion patterns.
- 48
- 46
- Unsupervised Multi-Domain Image Translation with Domain-Specific Encoders/Decoders44
This work proposes a novel and unified framework named Domain-Bank, which consists of a globally shared auto-encoder and domain-specific encoders/decoders, assuming that there is a universal shared-latent space can be projected and shows the comparable (or even better) image translation results over state-of-the-arts on various challenging unsupervised image translation tasks.
- BD-YOLOv8s: enhancing bridge defect detection with multidimensional attention and precision reconstruction40
BD-YOLOv8s, an advanced methodology utilizing YOLOv8s for bridge defect detection augments the network’s adaptability to a broad spectrum of bridge defect images through the integration of ODConv into the second convolutional layer, processing information within a four-dimensional kernel space.
- YOLOv8-MPEB small target detection algorithm based on UAV images36
This study proposes an algorithm for small target detection in UAV images based on an enhanced YOLOv8 model termed YOLOv8-MPEB that outperforms other prevalent detection algorithms in terms of accuracy and speed.
- 34
- Cascade Prompt Learning for Vision-Language Model Adaptation32
A novel Cascade Prompt Learning CasPL framework to enable prompt learning to serve both generic and specific expertise simultaneously, which achieves a significantly better balance between performance and inference speed, which is especially beneficial for deploying smaller VLM models in resource-constrained environments.
- MSGSE-Net: Multi-scale guided squeeze-and-excitation network for subcortical brain structure segmentation31
This paper proposes the multi-scale guided squeeze-and-excitation (MSGSE) attention module which can progressively and selectively aggregate discriminative features and proposes a novel entropy-weighted Dice loss (EDL) to force the network to focus on the ambiguous voxels around the boundaries of subcortical structures.
- 28
- 27
- Examining collaborative filtering algorithms for clothing recommendation in e-commerce26
The experimental results reveal that the recommendation based on user frequency similarity is very much suitable for the online clothing recommendation; the co-occurrence matrix reduction method is effective to improve the recommendation performance.
- 25
- Learnable differencing center for nighttime depth perception24
The key idea is to use recurrent inter-convolution differencing (RICD) and illumination affinitive intra-convolution differencing (IAICD) to enhance the nighttime color images and reduce the negative effects of the varying illumination, respectively.
- RigNet++: Semantic Assisted Repetitive Image Guided Network for Depth Completion23
This work designs a dense repetitive hourglass network (DRHN) to extract discriminative image features of complex environments, which can provide powerful contextual instruction for depth prediction and proposes a region-aware spatial propagation network (RASPN) for further depth refinement based on the semantic prior constraint.
- 22
- A more compact object detector head network with feature enhancement and relational reasoning21
This work proposes a more compact object detector head network (CODH), which can not only preserve global context information and condense the information density, but also allows instance-wise feature enhancement and relational reasoning in a larger matrix space.
- 20
- Understanding the Disharmony between Weight Normalization Family and Weight Decay20
This paper theoretically prove that ½λ∥W∥2 improves optimization only by modulating the effective learning rate and fairly has no influence on generalization when the weight normalization family is compositely employed.
- Non-Aligned Supervision for Real Image Dehazing18
This paper proposes an innovative dehazing framework that operates under non-aligned supervision, and introduces a self-attention network based on mean and variance for modeling real infinite airlight, using the dark channel prior as positional guidance.
- 17
- 14
- 13
- Simultaneous functional and structural imaging for photovoltaic devices13
An optical technique that can simultaneously acquire functionality and structure images of a single or multiple photovoltaic devices, simple, low-cost, effective, and efficient, and allows not only users to monitor the 'health condition' of the utilized devices, but also manufacturers to improve their fabrication conditions.
- Inter-Class Angular Loss for Convolutional Neural Networks13
Thorough experimental results on a series of vision and nonvision datasets confirm that ICAL critically improves the discriminative ability of various representative deep neural networks and generates superior performance to the original networks with conventional softmax loss.
- Zone Evaluation: Revealing Spatial Bias in Object Detection12
A new zone evaluation protocol is presented, extending from the traditional evaluation to a more generalized one, which measures the detection performance over zones, yielding a series of Zone Precisions (ZPs), and finds that the key lies in the human-imperceptible divergence in data patterns between objects in different zones, thus eventually forming a visible performance gap between the zones.
- Creative Birds: Self-Supervised Single-View 3D Style Transfer12
A novel method for single-view 3D style transfer that generates a unique 3D object with both shape and texture transfer that achieves state-of-the-art performance on the single-view 3D style transfer task is proposed.
- 12
- 11
- Towards more reliable evaluation in pedestrian detection by rethinking “ignore regions”10
A new pedestrian detection pipeline is devised, which considers the additional visible box at both the detection and evaluation stages, and is applied to dozens of detection methods and validated in reducing the over-reporting of false positives and providing more reliable evaluation results.
- 9
- 8
- 8
- 8
- YOLO11-SPE: a lightweight object detection model for corn seedling counting7
This paper proposes a lightweight corn seedling detection model, YOLO11-SPE, and constructs a drone image dataset for evaluation, based on diverse scenes and a specialized occlusion subset, which is sufficient to meet real-time detection requirements.
- 7
- 7
- 7
- 6
- MoCoLSK: Modality-Conditioned High-Resolution Downscaling for Land Surface Temperature6
The modality-conditioned large selective kernel (MoCoLSK) network is proposed, a novel architecture that dynamically fuses multimodal data through modality-conditioned projections and achieves a confluence of dynamic receptive field adjustment and multimodal feature fusion, leading to enhanced LST prediction accuracy.
- Design and Implementation of Equipment Maintenance Predictive Model Based on Machine Learning6
The service life of equipment is taken as the research object, machine learning theory is applied, the equipment maintenance prediction model based on random forest regression is built, and the typical technology implementation and verification of the model is given, which has good adaptability and reference value.
- 5
- 5
- A single-stream adaptive scene layout modeling method for scene recognition5
This paper proposes a single-stream adaptive scene layout modeling approach based on a layout modeling module (LMM), which constructs layouts without additional detection streams and adaptively captures the relations to take advantage of graph attention network.
- 5
- 5
- Face Recognition Using KPCA and KFDA5
The method of KPCA and KFDA is analyzed and their nature is revealed and the effectiveness of the algorithm is verified using the ORL database.
- EG-SpikeFormer: Eye-Gaze Guided Transformer on Spiking Neural Networks for Medical Image Analysis4
EG-SpikeFormer is introduced, an SNN architecture tailored for clinical tasks that incorporates eye-gaze data to guide the model's attention to the diagnostically relevant regions in medical images that demonstrates superior energy efficiency and performance in medical image prediction tasks.
- A Simple Detector with Frame Dynamics is a Strong Tracker4
This work proposes a simple yet effective infrared tiny-object tracker that enhances tracking performance by integrating global detection and motion-aware learning with temporal priors, and proposes a trajectory constraint filtering strategy in the post-processing stage to suppress false positives and enhance tracking robustness.
- Research on Detection and Recognition of Traffic Signs Based on Convolutional Neural Networks4
This paper focuses on the research of TSDR system and proposes a method based on Faster R-CNN combined with part of the low-level feature map that has a certain effect on the detection of small-scale traffic signs, and the detection speed is fast.
- 4
- Bridging Unpaired Facial Photos and Sketches by Line-Drawings4
A novel method to learn face sketch synthesis models by using unpaired data, termed sRender, which can generate multi-style sketches, and significantly outperforms existing unpaired image-to-image translation methods.
- Vaccine-YOLOv10: real-time QR code detection model for complex light condition3
Vaccine-YOLOv10 (VCY) is proposed to enhance QR code detection capability in complex light condition by introducing GhostConv and FasterC2f and Simulative Data Augment Algorithm (SDA) to simulate 5 types of complex light condition.
- Boosting Detection Accuracy: An Enhanced YOLOv8 for Small Target Detection in Remote Sensing3
The results suggest that the proposed YOLOv8s-Improved method addresses the challenges of small target detection in remote sensing and exhibits generalization capabilities across different datasets.
- A survey of historical learning: learning models with learning history3
This survey is the first survey that systematically studies the methodologies which make use of various historical statistics when training deep neural networks and encourages the community to pay attention to the think of historical learning principles when designing algorithms.
- 3
- 3
- Completing Visual Objects via Bridging Generation and Segmentation3
The method, named MaskComp, delineates the completion process through iterative stages of generation and segmentation, and demonstrates that the combination of one generation and one segmentation stage effectively functions as a mask denoiser.
- Discrete States-Based Trajectory Planning for Nonholonomic Robots3
Simulation experiments compared with prior works have validated that the DSTP algorithm has an order-of-magnitude efficiency advantage and can generate a smoother trajectory with a high speed and low control effort.
- Teaching-Learning-Feedback-Based Optimization3
Experimental results show that the proposed Teaching-learning-feedback-based Optimization has a superior and competitive capability in solving continuous optimisation problems.
- Evaluating Parameter Update Strategies in Deep Reinforcement Learning for Intelligent Traffic Signal Control2
Results show that dual-network decoupling accelerates training and raises network capacity, while the Dueling DDQN structure, by disentangling state value from advantages, curbs policy oscillation in multi-agent settings.
- 2
- Category-Aware 3D Object Composition with Disentangled Texture and Shape Multi-view Diffusion2
This paper proposes a straightforward yet powerful approach, category+3D-to-3D (C33D), for generating novel and structurally coherent 3D models, and introduces texture multi-view diffusion, which refines the textures of the remaining multi-view RGB images based on the novel 2D object.
- 2
- 2
- 2
- 1
- 1
- 1
- 1
- 1
- Global deformation model for 3D facial combination1
An automatic face and head deformation method that combines 3D faces with an arbitrary head model which provides an attractive solution for graphic design, Virtual Reality, 3D printing, and other industries, which can be widely used in consumer scenes.
- 1
- 1
- A lightweight YOLOv11n-based framework for accurate maize seedling detection and counting in field environments–
This study develops a lightweight maize seedling detection and counting model for field environments, with three key improvements based on the YOLOv11n architecture, providing an efficient and practical solution for automatic maize seedling detection and counting in field scenarios.
- –
- Unsupervised SAR Image Change Detection Method Based on Fused Difference Images–
A fusion method which combines the Gaussian weighted function and local energy and the convolutional wavelet neural network are proposed to generate a better DI for SAR image change detection.
- A region growing algorithm based on adaptive normal vector angle threshold using conditional entropy–
Experimental verification in the actual roadway environment shows that this method can effectively improve the segmentation accuracy of the roadway wall area, significantly enhance the automation and accuracy of the ranging process, and greatly improve the convenience of intelligent tunneling.
- Reinforcement Learning-Based Optimization of Quality-of-Service and Path Planning for Multi-UAV Mobile Edge Computing–
A unified reinforcement learning framework for a multi-UAV MEC platform that enhances Quality-of-Service (QoS) and optimizes path planning and the impact of sigmoid parameter tuning on risk-QoS trade-offs is highlighted, validating the adaptability of the proposed approach in complex scenarios.
- –
- Generative adversarial network design based on a dual-path snapshot spectral imaging system–
This study designed a generative adversarial network (GAN) comprising a generator and two discriminators specifically designed to correspond to the two physical measurements from the dual-path snapshot spectral system, thereby better assisting the generator in recovering the 3D spectral data cube from the aliased spectral image.
- –
- –
- –
- –
- A Generation Adversarial Network Based on Multi-Condition Confrontation and Gradient Optimization–
This paper proposes a multi-generator-based generation confrontation network improved multi-Generator generative adversarial nets (IMGAN), which introduces a hyperparameter to solve the disparity problem caused by multiple loss functions, and avoids excessive bias toward one of the gradient directions.
- –
- –
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.