A5047505143- Advanced Image and Video Retrieval Techniques
- Chaos-based Image/Signal Encryption
- Advanced MRI Techniques and Applications
- Scheduling and Optimization Algorithms
- Cerebrovascular and Carotid Artery Diseases
- Face and Expression Recognition
- Image Retrieval and Classification Techniques
- MRI in cancer diagnosis
- Advanced Vision and Imaging
- Advanced Image Processing Techniques
- Computer Graphics and Visualization Techniques
- Handwritten Text Recognition Techniques
- 3D Shape Modeling and Analysis
- Advanced Steganography and Watermarking Techniques
- Human Pose and Action Recognition
- Advanced Manufacturing and Logistics Optimization
- Sparse and Compressive Sensing Techniques
- Generative Adversarial Networks and Image Synthesis
- Smart Agriculture and AI
- AI in cancer detection
- Soft Robotics and Applications
- Optimization and Packing Problems
- Digital Media Forensic Detection
- Vehicle License Plate Recognition
- Advanced Neural Network Applications
Huazhong University of Science and Technology
2021-2024
Donghua University
2016-2024
Chengdu University of Traditional Chinese Medicine
2024
Nanjing University of Posts and Telecommunications
2024
Southern Medical University
2023-2024
Zhujiang Hospital
2024
NorthShore University HealthSystem
2020-2023
Guangxi Medical University
2022-2023
Kyushu University
2011-2023
University of Electronic Science and Technology of China
2023
This paper proposes a compression-combined digital image encryption method which is robust against consecutive packet loss and malicious shear attack. We utilize the characteristics of compressive sensing, dimensional reduction random projection, to compress encrypt simultaneously. On this basis, block Arnold scrambling used permutate position measurements. Bit wise XOR operation executed on quantized binary bit stream dissipate Gaussian distribution property cipher image. In paper, one...
Glabridin (Gla) has been reported to have significant effects in scar treatment, and however, the water insolubility of Gla leads its poor transdermal absorption ability, which affects bioactivities. Therefore, we attempted prepare dissolving microneedles (Gla-MN) improve absorbtion Gla. After investigation 3 factors including needle tip matrix concentration, prescription concentration backing material, dissolution method Gla, finally determined process parameters 10% hyaluronic acid (HA) as...
Purpose To probe the feasibility of deep learning‐based super‐resolution (SR) reconstruction applied to nonenhanced MR angiography (MRA) head and neck. Methods High‐resolution 3D thin‐slab stack‐of‐stars quiescent interval slice‐selective (QISS) MRA neck was obtained in eight subjects (seven healthy volunteers, one patient) at 3T. The spatial resolution high‐resolution ground‐truth data slice‐encoding direction reduced by factors 2 6. Four neural network (DNN) SR reconstructions were...
Accurately detecting and segmenting grape cluster in the field is fundamental for precision viticulture. In this paper, a new backbone network, ResNet50-FPN-ED, was proposed to improve Mask R-CNN instance segmentation so that detection performance can be improved under complex environments, shape variations, leaf shading, trunk occlusion, grapes overlapping. An Efficient Channel Attention (ECA) mechanism first introduced network correct extracted features better detection. To obtain detailed...
Peritoneal recurrence is the most frequent and lethal pattern in gastric cancer (GC) with serosal invasion after radical surgery. However, current evaluation methods are not adequate for predicting peritoneal GC invasion. Emerging evidence shows that pathomics analyses could be advantageous risk stratification outcome prediction. Herein, we propose a signature composed of multiple features extracted from digital hematoxylin eosin-stained images. We found was significantly associated...
This article presents a back-drivable spherical wrist with smart compliance for rendering human wrist-like capabilities in robotic applications. In order to synthetically impose both motion accuracy and active dedicated manipulations, robust hybrid control method is proposed that can switch between (at free motion) impedance interactions the environment). A sliding manifold formulated so when law drives zero, motion/impedance errors will automatically converge bounded regions despite system...
This study considers the batching and scheduling problem in two-stage hybrid flow shops which each job with a distinct due-date is processed through two serial production stages, of has identical machines parallel. Under fundamental trade-off that large batch sizes less frequent changeovers may reduce setup costs hence increase machine utilisation, while small times improve performance, to determine number batches, compositions, allocation batches parallel at stage, sequence allocated for...
Purpose To develop and test the feasibility of a sub‐3‐minute imaging strategy for non‐contrast evaluation extracranial carotid arteries using ungated quiescent interval slice‐selective (QISS) MRA, combining single‐shot radial sampling with deep neural network–based image processing to optimize quality. Methods The 12 human subjects were imaged at 3 T QISS MRA. In 7 healthy volunteers, effects Cartesian k‐space sampling, multishot acquisition (1.1‐3.3 seconds/slice, 141‐423 seconds/volume),...
Land use reflects the interrelation between physically built environment and activity patterns of people. It is indispensable information for decision-makes, but up-to-date accurate land often absent. Unlike approaches that make remote sensing data, in this work, we are interested a novel data source, tweets, explore its potential classification urban areas. Specifically, propose general framework classifying settlement land-use types by extracting location, time, quantity text features...
Edge is a type of valuable clues for scene character detection task. Generally, the existing edge-based methods rely on assumption straight text line to prune away non-character candidates. This paper proposes new method, called edge-ray filter, detect character. The main contribution proposed method lies in filtering out complex backgrounds by fully utilizing essential spatial layout edges instead line. Edges are extracted combination Canny and Preserving Smoothing Filter (EPSF). To...
Purpose This paper aims to present a bio-inspired neural network for improvement of information processing capability the existing artificial networks. Design/methodology/approach In network, authors introduce property often found in biological system – hysteresis as neuron activation function and bionic algorithm extreme learning machine (ELM) scheme. The give gradient descent procedure optimize parameters hysteretic develop an online select ELM parameters, including number hidden-layer...
Accurate 3D human models are useful for many applications in virtual fitting, ergonomics, film and television, video games. However, due to the limitations of scanners body privacy, creating a that accurately represents specific is challenging. Therefore, reshaping bodies based on anthropometric measurements has received extensive attention. existing methods have some drawbacks, such as inability reshaped change its posture, lack good link between real measurements, unreasonable...
Coupled Map Lattice (CML) usually serves as a pseudo-random number generator for encrypting digital images. Based on our analysis, the existing CML-based systems still suffer from problems like limited parameter space and local chaotic behavior. In this paper, we propose novel intermittent jumping CML system based multiple maps. The mechanism seeks to incorporate multi-chaos, dynamically switch coupling states relations, varying with spatiotemporal indices. Extensive numerical simulations...
Restricted preemption plays a crucial role in reducing total completion time while controlling overhead. A typical version of restricted preemptive models is k -restricted scheduling, where only allowed after task has been continuously processed for at least units time. Though solving this problem minimizing the makespan on parallel machines NP-hard general, it vital importance to obtain optimal solution small-sized problems, as well evaluation heuristics. This paper proposes strategies...
Reading life log is a type of techniques to automatically and unconsciously record people's reading intentions, interests habits. Besides, it can also serve as various assistants in our daily life. In this paper, reading-life system implemented by head-mounted unobtrusive video camera with high resolution shutter speed. We utilize DP matching, propose text-based frame mosaicing method integrate multiple frames clip. The developed tested the environments indoor outdoor. experimental results...