A5017318126- Advanced Neural Network Applications
- Advanced Image and Video Retrieval Techniques
- Domain Adaptation and Few-Shot Learning
- Gaze Tracking and Assistive Technology
- EEG and Brain-Computer Interfaces
- Video Surveillance and Tracking Methods
- Heat Transfer and Optimization
- Blind Source Separation Techniques
- UAV Applications and Optimization
- Spacecraft and Cryogenic Technologies
- Heat Transfer and Boiling Studies
- Computational Drug Discovery Methods
- Human Pose and Action Recognition
- Machine Learning in Materials Science
- 3D Printing in Biomedical Research
- Leaf Properties and Growth Measurement
- Refrigeration and Air Conditioning Technologies
- Surface Modification and Superhydrophobicity
- Visual Attention and Saliency Detection
- Laser-Plasma Interactions and Diagnostics
- RNA modifications and cancer
- Industrial Vision Systems and Defect Detection
- Phytochemistry and biological activity of medicinal plants
- Network Security and Intrusion Detection
- Chaos-based Image/Signal Encryption
Jiangsu University
2023-2025
Shenyang Jianzhu University
2024
University of Chinese Academy of Sciences
2018-2023
Waseda University
2023
Northeast Normal University
2020-2023
Hengyang Normal University
2019
Shanghai University
2015-2018
Kunming University
2013
Chinese Academy of Sciences
2009
In object detection, keypoint-based approaches often experience the drawback of a large number incorrect bounding boxes, arguably due to lack an additional assessment inside cropped regions. This paper presents efficient solution that explores visual patterns within individual regions with minimal costs. We build our framework upon representative one-stage detector named CornerNet. Our approach, CenterNet, detects each as triplet, rather than pair, keypoints, which improves both precision...
There are two mainstream approaches for object detection: top-down and bottom-up. The state-of-the-art mainly methods. In this paper, we demonstrate that bottom-up show competitive performance compared with have higher recall rates. Our approach, named CenterNet, detects each as a triplet of keypoints (top-left bottom-right corners the center keypoint). We first group according to some designed cues confirm locations based on keypoints. corner allow approach detect objects various scales...
In object detection, keypoint-based approaches often suffer a large number of incorrect bounding boxes, arguably due to the lack an additional look into cropped regions. This paper presents efficient solution which explores visual patterns within each region with minimal costs. We build our framework upon representative one-stage detector named CornerNet. Our approach, CenterNet, detects as triplet, rather than pair, keypoints, improves both precision and recall. Accordingly, we design two...
Detecting small objects is a challenging task due to their low resolution and noisy representation even using deep learning methods. In this paper, we propose novel object detection method based on the channel-aware deconvolutional network (CADNet) for accurate detection. Specifically, develop deconvolution (ChaDeConv) layer exploit correlations of feature maps in different channels across deeper layers, improving recall rate at additional computational costs. Following ChaDeConv layer,...
Object detection, instance segmentation, and pose estimation are popular visual recognition tasks which require localizing the object by internal or boundary landmarks. This paper summarizes these as location-sensitive proposes a unified solution named network (LSNet). Based on deep neural backbone, LSNet predicts an anchor point set of landmarks together define shape target object. The key to optimizing lies in ability fitting various scales, for we design novel loss function cross-IOU that...
Download This Paper Open PDF in Browser Add to My Library Share: Permalink Using these links will ensure access this page indefinitely Copy URL DOI
To address the limitations of traditional manual measurements phenotypes, a study focused on rapid acquisition phenotypic parameters in wheat grains was conducted. This research introduced an index for bud-point determination, which offers valuable insights into breeding selection, optimizing sowing and growth direction grains, assessing seed vitality. A grain collection device utilizing linear array camera designed constructed, accompanied by custom software to simplify operation. Image...
Although metastasis-associated lung adenocarcinoma transcript (MALAT)-1 is known to be consistently upregulated in several epithelial malignancies, little about its function or regulation. We therefore examined the relationship between MALAT-1 expression and candidate modulators such as DNA tumor virus oncoproteins human papillomavirus (HPV)-16 E6 E7, BK T antigen (BKVTAg), mouse polyoma middle (MPVmTAg) suppressor genes p53 pRb. Using suppressive subtractive hybridization (SSH) real-time...
The goal of object detection is to determine the class and location objects in an image. This paper proposes a novel anchor-free, two-stage framework which first extracts number proposals by finding potential corner keypoint combinations then assigns label each proposal standalone classification stage. We demonstrate that these two stages are effective solutions for improving recall precision, respectively, they can be integrated into end-to-end network. Our approach, dubbed Corner Proposal...
Information security has become increasingly important with the rapid development of mobile devices and internet.An efficient encryption system is a key to this end.In paper, we propose an image method based on cross diffusion two chaotic maps.We use sequences, namely Logistic map Chebyshev map, for generation which larger space than single one.Moreover, these sequences further decreases correlation neighboring pixels significantly.We conduct extensive experiments several well-known images...
With the advantage of high mobility, Unmanned Aerial Vehicles (UAVs) are used to fuel numerous important applications in computer vision, delivering more efficiency and convenience than surveillance cameras with fixed camera angle, scale view. However, very limited UAV datasets proposed, they focus only on a specific task such as visual tracking or object detection relatively constrained scenarios. Consequently, it is great importance develop an unconstrained benchmark boost related...
This paper develops a BCI simulated application system based on Unity3D, called going to the cinema. Firstly, 3D car model and some scene models built in 3ds Max are imported into Unity3D. Then, virtual reality layout is designed with its own models. The audio effects collision detection added enhance realistic immersion. After that, behaviors of defined. communication timer functions configured. Finally, can receive commands through TCP/IP protocol control movement go destination-cinema....