A5100446914- Medical Image Segmentation Techniques
- Face recognition and analysis
- Gait Recognition and Analysis
- Robotics and Sensor-Based Localization
- Brain Tumor Detection and Classification
- Human Pose and Action Recognition
- Anomaly Detection Techniques and Applications
- AI in cancer detection
- Advanced Image and Video Retrieval Techniques
- Biometric Identification and Security
- Advanced Vision and Imaging
- Digital Media Forensic Detection
- Advanced Graph Neural Networks
- Robotic Path Planning Algorithms
- Recommender Systems and Techniques
- Image Retrieval and Classification Techniques
- Image and Signal Denoising Methods
- User Authentication and Security Systems
- Mobile Agent-Based Network Management
- Image Processing and 3D Reconstruction
- Machine Fault Diagnosis Techniques
- Advanced Bandit Algorithms Research
- Advanced Steganography and Watermarking Techniques
- Corruption and Economic Development
- IPv6, Mobility, Handover, Networks, Security
Fuyang Normal University
2025
Hunan Normal University
2025
Harbin University of Science and Technology
2023-2024
Shanghai Polytechnic University
2024
Xidian University
2023
Minzu University of China
2023
Shanghai Jiao Tong University
2023
Shijiazhuang Tiedao University
2021-2023
Southeast University
2023
Shandong Academy of Sciences
2021
Automated segmenting and labeling of individual brain anatomical regions, in MRI are challenging, due to the issue structural variability. Although atlas-based segmentation has shown its potential for both tissue structure segmentation, inherent natural variability as well disease-related changes MR appearance, a single atlas image is often inappropriate represent full population datasets processed given neuroimaging study. As an alternative case use multiple atlases alongside label fusion...
Deepfake is a type of image and video face manipulation methods which could cause security society threats. Although some related databases detection models have been proposed for detecting forgery media, achieving generalizable detector both known unknown manipulations remains challenging. In this study, novel deepfake model with high generalizability to tackle issue. We employ supervised contrastive learning enhance the datasets. addition, we design cross-modality data augmentation method...
Cross-domain recommendation, as a cutting-edge technology to settle data sparsity and cold start problems, is gaining increasingly popular. Existing research paradigms primarily focus on leveraging the representation of overlapping entities, such aggregation or cross-domain consistency constraints, facilitate knowledge transfer enhance performance single-domain dual-domain recommender systems. Even though these approaches bring significant promotions, they still suffer from optimization...
Abstract Path planning for mobile robot aims to solve the problem of creating a collision‐free path from start state goal in given space, which is key supporting technology unmanned work. In order problems asymptotic optimal rapidly extended random trees star (RRT*) algorithm, such as its slow convergence rate, low efficiency and high cost path, an improved motion planner (Fast‐RRT*) was proposed based on hybrid sampling strategy choose parent backtracking. Firstly, bias constraint are...
The convenience of social media in communication and information dissemination has made it an ideal place for spreading rumor events, which raises a higher requirement automatic debunking events. Meanwhile, the traditional classification approaches relying on manual labeled features have to face daunting number efforts. In general, when facing dubious claim, people can authenticate verify realness event with contents continuous posts, such as source credibility, public sentiments,...
Exploring and mining subtle yet distinctive features between sub-categories with similar appearances is crucial for fine-grained visual categorization (FGVC). However, less effort has been devoted to assessing the quality of extracted representations. Intuitively, network may struggle capture discriminative from low-quality samples, which leads a significant decline in FGVC performance. To tackle this challenge, we propose weakly supervised Context-Semantic Quality Awareness Network...
Denial of service (DoS) attack can be easily launched for the simple discovery Protocol (SSDP) on Universal Plug and Play (UPnP). This paper proposes a novel DoS defense scheme UPnP. The connection between UPnP devices nodes will limited, useless descriptions received by control points filtered through setting packet-filter with traffic anomaly detection active rule-based XML firewall points. Finally, analyzes performance this scheme, which shows its particular advantages.
Aiming at the problem that it is difficult to determine shortest path of intelligent vehicle races, a optimization model for common sense proposed in this paper. Several appropriate discrete points are chosen on runway center line. According width and some constraints designed, GA algorithm used optimize along radius direction. In order further verify feasibility algorithm, race track established. Using method expatiated above, shorter obtained. Therefore effectiveness verified, purpose...
It is still a challenge to establish reliable links between transformation products (TPs) of the contaminant and microbiome. That because discovery TPs requires sophisticated experiment regime. Additionally, widely used linear approaches, for instance, Pearson's Spearman's correlations, may face challenges deal with this issue. due that 16S rRNA gene amplicon sequencing mass spectrometry datasets are compositional, which confound statistical inference. In study, we combined use 2H-labeled...
Since the Internet and social media grow rapidly in recent years. Image processing is playing an increasingly important role daily life. A key task for image super-resolution. In a super-resolution task, of high-resolution rate would be built on basis low-resolution picture. Super-resolution technologies could help recover damaged images recognize detail images. As neutral networks develop quickly, many industries have used deep learning to improve their efficiency. good performance...
It is still a challenge to establish reliable links between transformation products (TPs) of the contaminant and microbiome as discovery TPs requires sophisticated experiment regime. Additionally, linear approaches such Pearson’s Spearman’s correlations, which are widely used, may face challenges deal with this issue. due that 16S rRNA gene amplicon sequencing mass spectrometry datasets compositional, confound statistical inference. In study, we combined use 2H-labeled Stable...
In recent years, with the development of deep learning technology, neural networks have been widely used in field medical image segmentation. U-shaped Network(U-Net) is a segmentation network proposed for images based on full-convolution and gradually becoming most commonly architecture field. The encoder U-Net mainly to capture context information image, which plays an important role performance semantic algorithm. However, it unstable simple skip connection perform unstably global...