A5114488748- Medical Image Segmentation Techniques
- Image Retrieval and Classification Techniques
- Advanced Image and Video Retrieval Techniques
- Video Surveillance and Tracking Methods
- Face and Expression Recognition
- Face recognition and analysis
- Video Analysis and Summarization
- Human Pose and Action Recognition
- Advanced Steganography and Watermarking Techniques
- Image Processing and 3D Reconstruction
- Visual Attention and Saliency Detection
- Advanced Image Processing Techniques
- Image and Signal Denoising Methods
- Image Enhancement Techniques
- Advanced Neural Network Applications
- Olfactory and Sensory Function Studies
- Domain Adaptation and Few-Shot Learning
- Context-Aware Activity Recognition Systems
- Synthetic Aperture Radar (SAR) Applications and Techniques
- Handwritten Text Recognition Techniques
- Gaze Tracking and Assistive Technology
- Advanced Image Fusion Techniques
- Chaos-based Image/Signal Encryption
- Image Processing Techniques and Applications
- Medical Imaging Techniques and Applications
Guangdong Medical College
2022-2024
Prairie View A&M University
2023-2024
Dali University
2024
Yunnan Observatories
2024
Jilin Normal University
2023
National Supercomputing Center in Wuxi
2022
Asia University
2022
Jiangxi University of Science and Technology
2021
Wuhan University
2021
Guangzhou Vocational College of Science and Technology
2020
Transfer learning is extensively utilized for automatically recognizing and filtering out empty camera trap images that lack animal presence. Current research uses transfer identifying typically solely updates the fully connected layer of models, they usually select a pre-trained source model only based on its relevance to target task. However, do not consider optimization update selection, nor investigate effect sample size class number domain data set used construct performance model. Both...
In this paper, a robust algorithm of digital image watermarking based on discrete wavelet transform is introduced. It adds binary watermark into gray and the host required for detecting. The have two characteristics: (1) adapted to various images owing using mean value selected coefficients. (2) Arnold pretreatment eliminates spatial correlation disperses error bits among all pixels make more strongly against cropping operation. Experiment results show that proposed invisible common processing
Relevance feedback (RF) is an interactive process in content-based image retrieval (CBIR), which refines the retrievals to a particular query by using user's on previously retrieved results. In this paper, changing of relevance into learning problem neural network, technique for images network (NELIR) introduced, can improve user interaction with systems fully exploiting similarity information. NELIR describe distribution positive sample feature space set neighboring clusters produced...
Smart-phones are becoming our constant companions, they with us all of the time, being used for calling, web surfing, apps, music listening, TV viewing, social networking, buying, gaming, and a myriad other uses. technology that knows much better than most could imagine. Based on usage fact we never far away from smart phones, know where go, who interact with, what information consume, little clever software, can doing even why it. They beginning to ourselves. In this work present SenseSeer...
Because the data volume of news videos is increasing exponentially, a way to quickly browse sketch video important in various applications, such as media, archives and publicity. This paper proposes summarization method based on SURF features an improved clustering algorithm, overcome defects existing algorithms that fail account for changes shot complexity. Firstly, we extracted from sequences matched between adjacent frames, then detected abrupt gradual boundaries by calculating similarity...
Intelligent video surveillance technology can make real-time monitor for scenes which uses computer vision and analytic methods, it intelligent analysis without human intervention. The present research situation both at home abroad is given in this paper as well key technologies including detection, classification, tracking, behavioral analysis, the advantages of surveillance. At last, we give some difficulties predict direction future.
In the current scenario, research perspective on esophageal cancer becomes severe, high-prognosis malignancy; poor prognosis is primarily attributed to fact that most tumors remain asymptomatic and unrelated before it grows through esophagus. Significant decreases in mortality from may require effective approaches detect nurse more patients at early, curable stages. A new Improved Empirical Wavelet Transform (IEWT) dependent feature extraction approach a consistent homology for diagnosis of...
The rise of the national fitness trend has led more and people to join ranks physical fitness, current education resources human have not been able keep pace with development sports. Due lack guidance, many beginners sports are standardized, their movements effective even cause injuries. Sports action recognition technology can make up for this realize guidance players. In paper, we build a movement posture model based on acceleration sensors, introduce Selectable Mode Vocoder (SMV)...
Visual separability between different objects in various image classification tasks is highly uneven. As a consequence, humans need levels of detailed descriptions to separate multi-granularity similarities. Meanwhile, deep networks, such as convolutional neural networks (C-NNs) have demonstrated great ability multilevel representations for an object. Unfortunately, existing methods with typically use the output last layer only feature train flat N-way classifiers, which fail fit character....
Subject service provides the academic libraries opportunities to enhance core competiveness and realize sustainable development.It is a prerequisite create supportive external environment guarantee its sound development.And what more important promote library's internal
Respiratory-induced tumor motion presents a critical challenge in lung cancer radiotherapy, potentially impacting treatment precision and efficacy. This study introduces an innovative, deep learning-based approach for real-time, markerless tracking utilizing orthogonal X-ray projection images. It incorporates three key components: (1) sophisticated data augmentation technique combining hybrid deformable model with 3D thin-plate spline transformation, (2) state-of-the-art Transformer-based...
Scale estimation and occlusion detection are challenging problems in object tracking. Most existing methods fail to handle scale variations complex video sequences. This paper presents a novel approach for robust with multi-feature integration tracking framework, extract histogram of oriented lines (HOL) feature color-naming (CN) enhance description ability the target appearance. The proposed works by learning discriminative correlation filters based on pyramid representation. We learn...