A5100688917- Medical Image Segmentation Techniques
- Advanced Algorithms and Applications
- Fluid Dynamics and Mixing
- Advanced Image and Video Retrieval Techniques
- Image and Signal Denoising Methods
- Advanced Sensor and Control Systems
- Soft Robotics and Applications
- Fault Detection and Control Systems
- Image Retrieval and Classification Techniques
- Robot Manipulation and Learning
- Fluid Dynamics and Turbulent Flows
- Robotics and Sensor-Based Localization
- Image Processing Techniques and Applications
- Teleoperation and Haptic Systems
- Fluid Dynamics Simulations and Interactions
- Advanced Computational Techniques and Applications
- Advanced Vision and Imaging
- Analog and Mixed-Signal Circuit Design
- AI in cancer detection
- Robotic Path Planning Algorithms
- Advanced DC-DC Converters
- Visual Attention and Saliency Detection
- Structural Health Monitoring Techniques
- Industrial Technology and Control Systems
- Sensor Technology and Measurement Systems
Universität Hamburg
2010-2024
Shenyang University of Chemical Technology
2008-2024
South China University of Technology
2006-2024
Analysis and Testing Centre
2024
Inner Mongolia University of Technology
2020-2024
North China University of Water Resources and Electric Power
2023
Nantong University
2021-2023
Nanjing University of Information Science and Technology
2008-2022
Iwate University
2021
Zhejiang Ocean University
2021
We present an image captioning framework that generates captions under a given topic. The topic candidates are extracted from the caption corpus. A image's topics then selected these by CNN-based multi-label classifier. input to generation model is image-topic pair, and output of image. For this purpose, cross-modal embedding method learned for images, topics, captions. In proposed framework, topic, caption, organized in hierarchical structure, which preserved space using order-embedding...
Background subtraction is important for detecting moving objects in videos. Currently, there are many approaches to performing background subtraction. However, they usually neglect the fact that images consist of different whose conditions may change frequently. In this paper, a novel hierarchical model proposed based on segmented images. It first segments into several regions by mean-shift algorithm. Then, model, which consists region models and pixel models, created. The kind approximate...
Objective: The present study was designed to investigate the molecular mechanism and biological roles of lncRNA brain-derived neurotrophic factor antisense (lncRNA BDNF-AS) in acute spinal cord injury (ASCI). Methods: rat model ASCI hypoxic cellular were established detect expression BDNF-AS, miR-130b-5p, PR (PRDI-BF1 RIZ) domain protein 5 (PRDM5) cleaved caspase 3 (c-caspase 3) using qRT-PCR western blot. Basso, Beattie Bresnahan (BBB) score carried out assess neurological function. Flow...
Download This Paper Open PDF in Browser Add to My Library Share: Permalink Using these links will ensure access this page indefinitely Copy URL DOI
Experiments showed that Hierarchical Task Network (HTN) planners are suitable to find solutions for nontrivial tasks in complex scenarios. Mobile service robots able execute actions which may constitute the basic building blocks achieve high-level goals. However, only few experiments demonstrate application of a general purpose deliberative planner domain mobile robots. One challenging problem arises from fact adaptive AI-based presume closed-world assumption (CWA) and therefore unable deal...
We propose a new saliency-guided method for generating supervoxels in 3D space. Rather than using an evenly distributed spatial seeding procedure, our uses visual saliency to guide the process of supervoxel generation. This results densely distributed, small, and precise salient regions which often contain objects, larger less that correspond background. Our approach largely improves quality resulting segmentation terms boundary recall under-segmentation error on publicly available benchmarks.
It is often a difficult task to accurately segment brain magnetic resonance (MR) images with intensity in-homogeneity and noise. This paper introduces novel level set method for simultaneous MR image segmentation inhomogeneity correction. To reduce the effect of noise, anisotropic spatial information, which can preserve more details edges corners, proposed by incorporating inner relationships among neighbor pixels. Then energy function uses multivariate Student's t-distribution fit...
In this paper, a biped robot system for dynamic walking is presented. It has two 2-degree-of-freedom (DOF) lightweight legs and 6-DOF hip. All the joint pulleys of are driven by motors that placed at hip using steel cables. Since all heavy mounted hip, remarkably low-mass beyond which guarantees low inertia during high speeds. Utilizing cable-amplification mechanisms, stiffness strength achieved, resulting in better control performance compared to conventional direct-driven methods....
Taking Shanghai Science and Technology Museum as the operation target, robot named skycleaner which is totally actuated by pneumatic cylinders sucked to glass walls with vacuum suckers presented. In order solve problems of lower stiffness nonlinear movement characteristic system, a method segment variable bang-bang controller proposed implement accurate control position servo system using principle pulse width-modulation (PWM). Testing results show that can effectively improve quality. This...
Wall cleaning and maintenance of high-rise buildings is becoming one the most appropriate fields for robotization due to its current low level automation lack uniform building structure. The last decade has seen an increasing interest in developing employing service robots glass walls which always very dangerous strenuous mid-air. In this paper we propose a new kind robot named sky cleaner totally actuated by pneumatic cylinders sucked wall with vacuum suckers. constraints designing are...
Feature analysis and classification detection of abnormal cells from images for pathological are an important issue the realization computer assisted disease diagnosis. This paper studies a method cervical squamous epithelial cells. Based on cytological standard expert diagnostic experience, expressive descriptors extracted according to morphology, color, texture features scales Further, quantificational related cytopathology derived as well, including morphological difference degree, cell...
Fluid flow is of great importance to the quality near-net-shape beam blanks. Particle image velocimetry (PIV) was employed measure fluid inside a full-scale water model for blank continuous casting mould from five sections under different submerged entry nozzles (SENs) in present work. The results show that jet single-port SEN leads too significant impact on solidifying shell region flange-centre and fillets, especially fillet outer arc. resultant horizontal velocities three-port SENs are...