A5008957721- Retinal Imaging and Analysis
- Industrial Vision Systems and Defect Detection
- Image Processing Techniques and Applications
- Optical measurement and interference techniques
- Glaucoma and retinal disorders
- Digital Imaging for Blood Diseases
- Medical Image Segmentation Techniques
- Advanced Neural Network Applications
- Retinal Diseases and Treatments
- COVID-19 diagnosis using AI
- Image Retrieval and Classification Techniques
- Advanced Graph Neural Networks
- Textile materials and evaluations
- Image and Object Detection Techniques
- Advanced Measurement and Detection Methods
- Image Enhancement Techniques
- Image and Signal Denoising Methods
- Video Surveillance and Tracking Methods
- Radiomics and Machine Learning in Medical Imaging
- Retinal and Optic Conditions
- Advanced Image Processing Techniques
- Advanced Measurement and Metrology Techniques
- Surface Roughness and Optical Measurements
- Complex Network Analysis Techniques
- Spectroscopy and Chemometric Analyses
Tiangong University
2016-2025
Hebei University of Technology
2003-2025
Tianjin University
2003-2023
West China Medical Center of Sichuan University
2023
Tianjin University of Technology
2016-2021
Nanchang Hangkong University
2019
Tianjin Special Equipment Supervision and Inspection Technology Research Institute
2010
Driving fatigue is a main factor caused the traffic accidents. Our faces contain lot of useful information, we can use state eyes to detect fatigue, but eye would be affected by wearing sunglasses. In this paper, solve above problems and make algorithm keep accuracy real-time at same time, infrared videos for detecting propose an recognition method based on convolution neural network (CNN), eventually calculating percentage eyelid closure over pupil time (PERCLOS), blink frequency fatigue....
Lung cancer has one of the highest morbidity and mortality rates in world. nodules are an early indicator lung cancer. Therefore, accurate detection image segmentation is great significance to diagnosis This paper proposes a CT (Computed Tomography) nodule method based on 3D-UNet Res2Net, establishes new convolutional neural network called 3D-Res2UNet. 3D-Res2Net symmetrical hierarchical connection with strong multi-scale feature extraction capabilities. It enables express features finer...
Lung cancer has become one of the life-threatening killers. disease need to be assisted by CT images taken doctor's diagnosis, and segmented image lung parenchyma is first step help doctor diagnosis. For problem accurately segmenting parenchyma, this paper proposes a segmentation method based on combination VGG-16 dilated convolution. First all, we use three parts network structure convolution pooling input image. Secondly, using multiple sets convolutions make large enough receptive field....
We address the problem of disentangled representation learning with independent latent factors in graph convolutional networks (GCNs). The current methods usually learn node by describing its neighborhood as a perceptual whole holistic manner while ignoring entanglement factors. However, real-world is formed complex interaction many (e.g., same hobby, education or work social network). While little effort has been made toward exploring GCNs. In this paper, we propose novel Independence...
Fatigue driving has become one of the major causes traffic accidents. The authors propose an effective method capable detecting fatigue state via spatial–temporal feature driver's eyes. In this work, consider detection as image‐based sequence recognition and end‐to‐end trainable convolutional neural network with long short‐term memory (LSTM) units is designed. First, apply a deep cascaded multi‐task framework to extract eye region from infrared videos. Then spatial features are learned by...
Rice grain moisture has a great impact on th production and storage quality of rice. The main objective this study was to design develop rapid-detection sensor for rice based the Near-infrared spectroscopy (NIR) characteristic band, aiming realize its accurate on-line measurement. In paper, NIR spectral information samples with different content obtained using portable spectrometer. Then, partial least squares (PLS) competitive adaptive reweighted (CARS) were applied model analyze data find...
Objective: Glaucoma is a leading cause of irreversible visual impairment and blindness worldwide, primarily linked to increased intraocular pressure (IOP). Early detection essential prevent further impairment, yet the manual diagnosis retinal fundus images (RFIs) both time-consuming inefficient. Although automated methods for glaucoma (GD) exist, they often rely on individual models with manually optimized hyperparameters. This study aims address these limitations by proposing an...
Graph contrastive learning (GCL), as a typical self-supervised paradigm, has been able to achieve promising performance without labels and gradually attracts much attention. Graph-level method aims learn representations of each graph by contrasting two augmented graphs. Previous studies usually simply apply keep the embeddings views from same anchor (positive pairs) close other, well separate different graphs (negative pairs). However, it is well-known that structure always complex...
ABSTRACT Facial expression plays a crucial role during interactions with people. Previous studies on facial recognition (FER) have mainly focused adults, while there are few FER for infants. Due to the apparent differences in proportions and contours between infants could not be conducted existing datasets. In order study infant expressions in‐depth, we create (IFER) dataset by collecting 10,240 images. Since infants' faces smooth lines weak sharpness, inter‐class similarity of is higher...
Non-proliferative diabetic retinopathy is the early stage of retinopathy. Automatic detection non-proliferative significant for clinical diagnosis, screening and course progression patients. This paper introduces design implementation an automatic system based on color fundus images. Firstly, structures, including blood vessels, optic disc macula, are extracted located, respectively. In particular, a new localization method using parabolic fitting proposed physiological structure...
Currently, lung cancer has one of the highest mortality rates because it is often caught too late. Therefore, early detection essential to reduce risk death. Pulmonary nodules are considered key indicators primary cancer. Developing an efficient and accurate computer-aided diagnosis system for pulmonary nodule important goal. Typically, a consists two parts: candidate extraction false-positive reduction nodules. The false positives (FPs) remains challenge due morphological characteristics...
Automated ultrasonic signal classification systems are often utilized for the recognition of a large number signals in engineering materials. Existing defect methods mainly image-based and serve to extract features various defects. In this paper, we propose novel detection baseline model based on fully convolution network (FCN) gated recurrent unit (GRU) classify from flawed 3D braided composite specimens with debonding proposed algorithm, Gated Recurrent Unit Fully Convolutional Network...