A5079480534- Image Processing Techniques and Applications
- Cell Image Analysis Techniques
- EEG and Brain-Computer Interfaces
- Muscle activation and electromyography studies
- AI in cancer detection
- Neuroscience and Neural Engineering
- Image Retrieval and Classification Techniques
- COVID-19 diagnosis using AI
- Radiomics and Machine Learning in Medical Imaging
- Medical Image Segmentation Techniques
- Advanced Neural Network Applications
- Infrared Thermography in Medicine
- Anomaly Detection Techniques and Applications
- Advanced Sensor and Energy Harvesting Materials
- ECG Monitoring and Analysis
- Hand Gesture Recognition Systems
- Digital Imaging for Blood Diseases
- Functional Brain Connectivity Studies
- Brain Tumor Detection and Classification
- Traffic Prediction and Management Techniques
- Cervical Cancer and HPV Research
- Vehicle License Plate Recognition
- Handwritten Text Recognition Techniques
- Stroke Rehabilitation and Recovery
- Digital Media Forensic Detection
Shenzhen Institutes of Advanced Technology
2022-2025
Chinese Academy of Sciences
2022-2025
University of Chinese Academy of Sciences
2022-2024
Shenzhen Academy of Robotics
2024
Northeastern University
2019-2022
Institute of Electrical and Electronics Engineers
2022
Ministry of Education of the People's Republic of China
2020
BACKGROUND: The novel coronavirus disease 2019 (COVID-19) constitutes a public health emergency globally. number of infected people and deaths are proliferating every day, which is putting tremendous pressure on our social healthcare system. Rapid detection COVID-19 cases significant step to fight against this virus as well release off the OBJECTIVE: One critical factors behind rapid spread pandemic lengthy clinical testing time. imaging tool, such Chest X-ray (CXR), can speed up...
Microorganisms play a great role in ecosystem, wastewater treatment, monitoring of environmental changes, and decomposition waste materials. However, some them are harmful to humans animals such as tuberculosis bacteria plasmodium. In course, it is important identify, track, analyze, consider the beneficial side get rid negative effects microorganisms using fast, accurate, reliable methods. recent decades, image analysis techniques have been used address drawbacks manual traditional...
The main obstacle to image augmentation with Generative Adversarial Networks (GANs) is the need for a large amount of training data, but this difficult small datasets like Environmental Microorganisms (EMs). EM analysis plays vital role in environmental monitoring and protection, it often encountered due difficulty collection. To end, we propose an Enhanced Framework GANs (EF-GANs) that combines geometric transformation methods augmentation. First all, color has insignificant impact on its...
In Autonomous Vehicle (AV) technology, the robustness of Traffic Sign Classification (TSC) systems is crucial for ensuring safe navigation. Currently, TSC lack dedicated adversarial recovery methods, making them susceptible to attacks. This study embarks on an innovative path by adapting six established methods from general image classification (IC) in AVs. Prompted limited availability TSC-specific solutions, our research undertakes a comparative analysis these evaluate their applicability...
This paper explores the integration of Artificial Intelligence Generated Content (AIGC) with human-machine intelligence (HMI) to enhance functionality Intelligent Transportation Systems (ITS). Adaptive decision-making mechanisms are crucial as transportation networks become increasingly complex, generating vast real-time data from vehicles, infrastructure, and users. AIGC plays a transformative role in optimizing traffic flow through dynamic routing management. At same time, human ensures...
Environmental Microorganism Data Set Fifth Version (EMDS-5) is a microscopic image dataset including original (EM) images and two sets of Ground Truth (GT) images. The GT include single-object set multi-object set. EMDS-5 has 21 types EMs, each which contains 20 EM images, can realize to evaluate preprocessing, segmentation, feature extraction, classification retrieval functions. In order prove the effectiveness EMDS-5, for function, we select most representative algorithms price indicators...
In this Letter, a hierarchical conditional random field (HCRF) model‐based gastric histopathology image segmentation (GHIS) method is proposed, which can localise abnormal (cancer) regions in images to assist histopathologists medical work. First, obtain pixel‐level information, the authors retrain convolutional neural network (CNN) build up their potentials. Then, abundant spatial information patch level, they fine tune another three CNNs patch‐level Thirdly, based on pixel‐ and potentials,...
A behavior description helps analyze tiny objects, similar objects with weak visual information, and information. It plays a fundamental role in the identification classification of dynamic microscopic videos. To this end, we propose foldover features to describe objects. Foldover is defined as: Each frame an object's motion superimposed on same spatial plane spacetime order motion, result superposition motion. object contains temporal static features. Therefore, extracted based are In work,...
To assist researchers to identify Environmental Microorganisms (EMs) effectively, a Multiscale CNN-CRF (MSCC) framework for the EM image segmentation is proposed in this paper. There are two parts framework: The first novel pixel-level approach, using newly introduced Convolutional Neural Network (CNN), namely, “mU-Net-B3”, with dense Conditional Random Field (CRF) postprocessing. second VGG-16 based patch-level method “buffer” strategy, which further improves quality of details EMs. In...
The control performance of myoelectric prostheses would not only depend on the feature extraction and classification algorithms but also interactions dynamic window-based hyperparameters (WBHP) used to construct input signals. However, relationship between these how they influence convolutional neural networks (CNNs) during motor intent decoding has been studied. Therefore, we investigated impact various combinations WBHP (window length overlap) employed for construction raw two-dimensional...
Prediabetes, characterized by elevated blood glucose (BG) levels without reaching the threshold for diabetes, necessitates early detection to avert complications. Unfortunately, traditional BG monitoring methods involve painful finger pricking. Hence, exploring noninvasive alternatives estimation and continuous is imperative. This paper investigates electroencephalogram (EEG) frequency parameters, an underexplored aspect of prediabetes diagnosis. Our investigation involved 25 participants...
The use of deep neural networks in electromyogram (EMG) based prostheses control provides a promising alternative to the hand-crafted features by automatically learning muscle activation patterns from EMG signals. Meanwhile, raw signals as input convolution (CNN) offers simple, fast, and ideal scheme for effective prostheses. Therefore, this study investigates relationship between window length overlap, which may influence generation robust 2-dimensional (2D) application CNN. And rule thumb...
Prediabetes is a metabolic disorder where the blood glucose (BG) level higher than normal but not high as diabetes; early diagnosis can prevent health complications and death. However, to determine BG level, it required prick finger, which causes pain discomfort. To eliminate this problem, there need investigate noninvasive techniques estimate values continuous monitoring. In paper, we investigated changes in electroencephalogram (EEG) frequency parameters that have been scarcely considered...
Surface electromyogram (sEMG) is arguably the most sought-after physiological signal with a broad spectrum of biomedical applications, especially in miniaturized rehabilitation robots such as multifunctional prostheses. The widespread use sEMG to drive pattern recognition (PR)-based control schemes primarily due its rich motor information content and non-invasiveness. Moreover, recordings exhibit non-linear non-uniformity properties inevitable interferences that distort intrinsic...
A major barrier to the commercialization of pattern recognition (PR)-based myoelectric prostheses is lack robustness confounding factors such as electrode shift which has been lingering for years. To overcome this challenge, a novel Duo-Stage Convolutional Neural Network (DS-CNN) proposed. The DS-CNN comprised two cascaded stages in first stage deciphers occurrence particular kind upon requisite CNN model triggered second accurate decoding individual motion intent, necessary initiating...