Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks76 from public data
- Implicit Sample Extension for Unsupervised Person Re-Identification144
An Implicit Sample Extension method to generate support samples from actual samples and their neighbouring clusters in the embedding space through a progressive linear interpolation (PLI) strategy, which reduces the “sub and mixed” clustering errors, thus improving the Re-ID performance.
- VisDrone-DET2018: The Vision Meets Drone Object Detection in Image Challenge Results133
A large-scale drone-based dataset, including 8, 599 images with rich annotations, including object bounding boxes, object categories, occlusion, truncation ratios, etc, is released, to narrow the gap between current object detection performance and the real-world requirements.
- VisDrone-SOT2018: The Vision Meets Drone Single-Object Tracking Challenge Results87
The evaluation protocol of the VisDrone-SOT2018 challenge is presented and the results of a comparison of 22 trackers on the benchmark dataset are presented, which are publicly available on the challenge website.
- 84
- 73
- 64
- 49
- Robust visual tracking with channel attention and focal loss35
This work introduces channel attention and focal loss into the network design to enhance feature representation learning and demonstrates that the enhanced tracker achieves significant performance improvement while running at a real-time frame-rate of 66 fps.
- 34
- Cross-ratio invariant based line scan camera geometric calibration with static linear data33
A flexible new technique to calibrate the geometric model of a line scan camera by using the properties of cross-ratio and point correspondences, which is easy to use and flexible.
- 26
- 23
- 22
- Drone-Based Visible–Thermal Object Detection with Transformers and Prompt Tuning18
This study introduces an innovative dual-modal object detection framework called Visual Prompt multi-modal Detection (VIP-Det) that harnesses the Transformer architecture as the primary feature extractor and integrates vision prompts for refined feature fusion.
- Masked and dynamic Siamese network for robust visual tracking18
Two modules are proposed, namely the target objectness model and the target template model, based on existing Siamese network based trackers to solve the distraction problem and the Gaussian mixed model enhances diversity and simultaneously reduces redundancy between target samples.
- Double HEVC Compression Detection with the Same QPs Based on the PU Numbers18
A novel method based on the same Qps is applied in dealing with double HEVC compression detection and it is proved that the proposed algorithm is of high classification accuracy for detecting doubleHEVC compression.
- 17
- Learning target-aware correlation filters for visual tracking17
The TACF formulation is proposed, which proposes an optimization strategy based on the Preconditioned Conjugate Gradient method for efficient filter learning and achieves state-of-the-art performance on OTB100 and the top rank in EAO on the VOT2016 challenge.
- Learning adaptively windowed correlation filters for robust tracking17
A generic DCF tracking framework which suppresses background information and highlights the foreground object with an object likelihood map computed from the color histograms is proposed which can be combined with many DCF trackers and achieves significant performance improvement.
- An Online Intelligent Method to Calibrate Radar and Camera Sensors for Data Fusing15
An intelligent method to calibrate radar and camera sensors for data fusing without human assistance operation in the general scenarios with low time cost or easy operation is introduced.
- 15
- 15
- Drone Based RGBT Tracking with Dual-Feature Aggregation Network13
A Dual-Feature Aggregation Network that utilizes attention mechanisms in both spatial and channel dimensions to aggregate heterogeneous modality feature information to better address the problem of visible-infrared information fusion is proposed.
- 13
- Self-Guided Hard Negative Generation for Unsupervised Person Re-Identification13
A joint framework is developed which incorporates a hard negative generation network (HNGN) and a re-ID network which is alternately trained in an adversarial manner to improve each other, where the reID network guides H NGN to generate challenging data and HNGN enforces the re-IDs to enhance discrimination ability.
- When correlation filters meet fully-convolutional Siamese networks for distractor-aware tracking13
This paper incorporates correlation filter trackers with the fully-convolutional Siamese network (SiamFC) to achieve distractor-aware tracking and combines the advantage of end-to-end tracking based on deep features in SiamFC and online model-update in correlation filterTrackers based on hand-crafted features.
- Multiple-Object-Tracking Algorithm Based on Dense Trajectory Voting in Aerial Videos12
A multiple-object-tracking algorithm based on dense-trajectory voting in aerial videos that could obtain good target-tracking results in various complex scenarios, and was still robust at a low frame rate by changing the video frame rate.
- Discontinuity-preserving optical flow algorithm12
A modification of Horn and Schunk's approach is investigated, which leads to a better preservation of flow discontinuities and introduces a more effcient iterative method than the Gauss –Seidel method to solve the associated Euler –Lagrange equation.
- Collaborative Convolution Operators for Real-Time Coarse-to-Fine Tracking11
This paper presents a coarse-to-fine tracking (CTFT) framework which breaks the task of visual tracking into two stages and maintains the efficient element-wise solution of standard DCF.
- 11
- Hyper-Siamese network for robust visual tracking10
This paper designs a skip-layer connection network named Hyper-Siamese to aggregate the hierarchical feature maps of SiamFC and constitute the hyper-feature representations of the target and proves that different convolutional layers are all useful for object tracking.
- 10
- Learning Fully Convolutional Network for Visual Tracking With Multi-Layer Feature Fusion8
This paper presents a new approach for general object tracking based on the fully convolutional network with multi-layer feature fusion, which outperforms many other state-of-the-art trackers.
- 7
- Spatio-temporally weighted multiple instance learning for visual tracking7
A spatio-temporally weighted MIL (STWMIL) tracker which integrates temporal weight into the update scheme for Haar-like features and spatial weight intoThe bag probability function is proposed using the weighted Noisy-OR model.
- Non-intrusive residential load identification based on load feature matrix and CBAM-BiLSTM algorithm6
A novel method of residential load identification based on load feature matrix and improved neural networks is proposed and the identification results are optimized with dynamic time warping (DTW).
- Beyond feature integration: a coarse-to-fine framework for cascade correlation tracking6
This paper presents a coarse-to-fine framework for cascade correlation tracking (CCT), which decomposes a complicated tracker into two low-complexity modules, a coarse tracker and a refined tracker working in a coarse-to-fine manner.
- 6
- 5
- 5
- Learning a Twofold Siamese Network for RGB-T Object Tracking5
A twofold Siamese network for RGB-T tracking, which is composed of an RGB branch and a thermal branch, which can be utilized to compute similarities between the search image and the exemplar image and two criteria are proposed to measure the confidence degrees of response maps obtained by these two modalities.
- 5
- Towards Unified Transformer for UAV‐Based Multi‐Task Oriented Object Detection4
TUMODet is proposed, a unified multi‐task oriented detection framework built upon the spatial transform decoupling (STD) detector, enabling a single model to dynamically handle three detection tasks (visible, thermal, and fused modalities).
- 4
- Object Tracking in Satellite Videos Based on a Lightweight Network4
A lightweight network is designed for the satellite video tracking task that generates a response map with the same size as the input image and reserves the spatial resolution of targets and transfer parameters are transferred from an existing network and trained with the initial annotated frame.
- 4
- 4
- 3
- 3
- Depth-Camera Calibration Optimization Method Based on Homography Matrix3
Compared with Zhang's calibration method, the relative errors of the focal length and the coordinates obtained by the proposed self-calibration calibration method are smaller, in addition it has a good restoration effect for the edge distortions of the depth images.
- 3
- 3D reconstruction with two webcams and a laser line projector3
A 3D reconstruction desktop system is built based on binocular stereo vision using a laser scanner based on epipolar rectification and stereo correspondences are established robustly from epipolar geometry and the laser shadow on the scanned object.
- 2
- An airborne object detection and location system based on deep inference2
This paper combines the improved object detection algorithm ST-YOLO based on YOLOX and Swin Transformer with the visual positioning algorithm and deploy it on the airborne end by using TensorRT to realize the detection and location of objects during the flight of the drone.
- 2
- LCO: Lightweight Convolution Operators for fast tracking2
This work tackles the problems of model complexity and over-fitting by introducing Lightweight Convolution Operators (LCO) in the Discriminative Correlation Filters (DCFs) based trackers, and reduces over 90% of the redundant trainable parameters in the tracking model.
- A new image enhancement method based-on FFT2
An inverse-gauss high pass filter and a low pass filter are designed to enhance image quality, which not only denoise but also reserve the fine information of image.
- 1
- 1
- 1
- 1
- Autocorrelated correlation filter for visual tracking1
This work proposes a rigorous solution to closed-form correlation filter tracking that can be computed pixel by pixel from a small linear equation system and achieves better tracking performance than the baseline element-wise solution.
- 1
- 1
- A fast super-resolution algorithm based on ideal sample1
A new fast super-resolution algorithm is proposed, which utilized the relation of high-resolution image and low- resolution image and phase shift relation of low-resolution images and PSNR is more 10dB than traditional interpolation algorithm.
- –
- –
- Gesture Recognition Based on Multi-Channel Convolutional Neural Network–
The multi-channel convolutional network of this paper is used to train and use the Multi-perspective Static Gesture Database constructed in this paper for experimental training and further research is carried out to compare the three-Perspective combination and the two perspective combinations respectively.
- –
- –
- –
- New super-resolution reconstructing algorithm in image frequency domain–
Three simulation experimental results indicate that the proposed algorithm removes the limits on frame numbers and sub-pixel shifts between frames and makes the PSNR of SR images to be increased by about 5dB.
- A discriminant analysis method for prokaryotic promoter regions based on feature selection–
A discriminant analysis method based on feature selection for prokaryotic promoter regions is proposed that outperforms most of the existing approaches based on several performance measurements.
- –
- Reconstruction of high-resolution image based on interpolation kernel–
A novel algorithm for reconstructing high resolution image based on 2D interpolation kernel by use of continuous image sampling and interpolation formula that makes common low-resolution sequence images relate to standard displacements and reconstructs high-resolution image according to the relationship between low- Resolution sequence images with standard displacement and high- resolution image.
- A Novel Algorithm for Finding Interspersed Repeat Regions–
A novel projection-assemble algorithm to find unknown interspersed repeats in DNA sequences, which found an un-labeled repeat region that occurs five times in Escherichia coli genome, with its length more than 5,000 bp, and a mismatch probability less than 4%.
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.