A5100379801- Video Surveillance and Tracking Methods
- Anomaly Detection Techniques and Applications
- Coal Properties and Utilization
- Robotic Path Planning Algorithms
- Advanced Algorithms and Applications
- Neural Networks and Applications
- Industrial Vision Systems and Defect Detection
- Advanced Sensor and Control Systems
- Face recognition and analysis
- Image Enhancement Techniques
- Autonomous Vehicle Technology and Safety
- Fault Detection and Control Systems
- Blockchain Technology Applications and Security
- Reinforcement Learning in Robotics
- Generative Adversarial Networks and Image Synthesis
- Digital Media Forensic Detection
- Image Retrieval and Classification Techniques
- Oil and Gas Production Techniques
- Non-Destructive Testing Techniques
- Granular flow and fluidized beds
- Time Series Analysis and Forecasting
- Text and Document Classification Technologies
- Geomechanics and Mining Engineering
- Advanced Malware Detection Techniques
- Image and Object Detection Techniques
Chalmers University of Technology
2022-2024
University of Sheffield
2021-2024
Harbin Institute of Technology
2022-2024
Heilongjiang Institute of Technology
2024
Xiangtan University
2024
Shanghai Electric (China)
2024
Salus (United States)
2023
Central China Normal University
2022-2023
Shanghai Tongji Urban Planning and Design Institute
2023
Tongji University
2023
With the increasing popularity and wide application of Internet, users (such as managers data consumers) in Industrial Internet Things (IIoT) can remotely analyze control real-time collected by various smart sensor devices. However, there are many security privacy issues process transmitting through public channels IIoT environment. In order to against illegal access opponents, a novel anonymous user authentication key agreement scheme based on hash elliptic curve encryption is proposed this...
This paper aims to explore the application of deep learning in smart contract vulnerabilities detection. Smart contracts are an essential part blockchain technology and crucial for developing decentralized applications. However, can cause financial losses system crashes. Static analysis tools frequently used detect contracts, but they often result false positives negatives because their high reliance on predefined rules lack semantic capabilities. Furthermore, these quickly become obsolete...
Existing face relighting methods often struggle with two problems: maintaining the local facial details of subject and accurately removing synthesizing shadows in relit image, especially hard shadows. We propose a novel deep method that addresses both problems. Our learns to predict ratio (quotient) image between source target desired lighting, allowing us relight while details. During training, our model also modify by using estimated shadow masks emphasize on high-contrast borders....
The existing precision grazing technology helps to improve the utilization rate of livestock pasture, but it is still at level “collectivization” and cannot provide more accurate management control. (1) Background: In recent years, with rapid development agent-related technologies such as deep learning, visual navigation tracking, lightweight edge computing cell target detection algorithms have been proposed. (2) Methods: this study, improved YOLOv5 detector combined extended dataset...
Monitoring of complex industrial processes can be achieved by obtaining process data utilising various sensing modalities. The recent emergence deep learning provides a new routine for processing multi-sensor information. However, the ability shallow neural networks is insufficient, and amount required often too large scenarios. This paper novel transfer method as possible solution that offers an advantage better network without requirement training data. presents how Transformer with...
Speller is the best way to express performance of brain-computer interface (BCI) paradigm. Due its advantages short analysis time and high accuracy, SSVEP paradigm has been widely used in BCI speller system based on wet electrode. It known that electrode operation cumbersome subjects have a poor experience. In addition, asynchronous threshold analysis, flickers continuously from beginning end experiment, which leads visual fatigue. The dry simple provides comfortable experience for subjects....
The shortage of annotated ECG data presents a significant impediment, hampering the overall generalization capabilities machine learning models tailored for automated classification. collective integration multisource datasets potential remedy this challenge. However, it is crucial to underscore that mere addition supplementary does not automatically guarantee performance enhancement, given unresolved challenges associated with data. In research, we address one such challenge, namely, issue...
This paper presents an in-depth study on the indirect drying characteristics and interface renewal. By comparing rate curve, it is found that shortening surface stage, which contains a repeated evaporation-diffusion-condensation process of moisture, key to improve total efficiency. stirring coal realizing renewal between bottom layer layer, efficiency reached seven times than static drying. Based heat transfer with high mass rate, new type moving bed moisture control device designed. At same...
Underwater light absorption and scattering lead to color deviation, low brightness, fuzzy details contrast of underwater images. In this contribution, an image enhancement algorithm based on balance multi-scale fusion is proposed. Firstly, a method used correct the color. Then, improved dark channel prior algorithm, local information balanced derive two atmospheric lights transmission maps, which are transformed into images with enhanced respectively. Finally, adopted fuse brightness...
Adsorption offers an efficient and eco-friendly method for the recovery of rare earth elements (REEs). However, it is not well known that most adsorbents are eco-friendly. Rational design next generation should aim to enhance their green properties. In this work, sepiolite, Fe3O4 itaconic acid were selected as raw materials. Poly(itaconic acid)/magnetite sepiolite composite (PIA/MSEP) was prepared using water solvent under mild reaction conditions entire process uses neither bulk use toxic...
Abstract A green revolution has accelerated over the recent decades with a look to replace existing transportation power solutions through adoption of greener electrical alternatives. In parallel digitisation manufacturing enabled progress in tracking and traceability processes improvements fault detection classification. This paper explores machine manufacture challenges faced identifying failures modes during this life cycle demonstration state-of-the-art vision methods for classification...
A modified Lattice–Boltzmann method is proposed by considering the Klinkenberg effect and adsorbability-desorbability for purpose of simulating methane gas seepage in fissured coal. The results show that has a little influence on coal, so it can be neglected engineering computations simplicity. If both are considered, pressure decreases as Darcy coefficient increases. It found drainage simulations near hole, adsorption desorption cannot neglected, location hole great efficient λ when just...
This paper discussed how we should apply fuzzy self-tuning PID controller to AC-speed adjustable system, and emphasized realize this using search table method in C++ language based on TMS320F2812, which is the newest 32-bit fixed point DSP from TI company. The author eventually applied PMSM-speed vector control system MCK2812, a motion development kit of technosoft And results show that can improve dynamic characteristics static obviously.
Induction motor (IM) is used in many modern industrial applications such as fans, blowers and pumps so on. Inrush current torque pulsation exist when IMs starts. IM terminal voltage reduced by three pairs of anti-parallel thristors to fed traditional soft starter. Therefore, a novel fuzzy PID strategy for the starting induction motors has been proposed paper. Based on this Simulink model starter implemented Matlab/Simulink. The results simulation experiments show that way practical.
The extraction of characteristic parameters is extremely important front-end the speech recognition system, accurate Mongolian have meaning to its technology development. This paper clarified Mel frequency cepstrum coefficients(MFCC) features and method at first, then it will treated with first-order second-order difference, after that, combining these get a new feature parameter vector. Analyze extract this eigenvector from signal by means MATLAB. Experimental results reveal that more is,...
Industrial control protocol feature extraction is an important way to improve the accuracy and speed of industrial traffic classification. This paper firstly proposes a keyword method for protocol, then designs implements system (ICS) classification based on this method. The proposed utilizes characteristics relatively fixed format periodicity in ICS. features can be accurately extracted after data preprocessing, segmentation, redundant filtering, byte mining. A dataset formed. designed ICS...
Pedestrian detection in a crowd is very challenging issue. This paper addresses this problem by novel NonMaximum Suppression (NMS) algorithm to better refine the bounding boxes given detectors. The contributions are threefold: (1) we propose strict-NMS, which applies higher suppression threshold an instance; (2) apply Mask boot module Fast-RCNN improve detect capacity. (3) We achieve state of art results on CityPersons benchmarks.
An auto-adaptive scheme of embedding watermark based on wavelet transformation domain is proposed in this paper. Firstly, the applies transform to original image, confirms ideal embedded location low-frequency coefficient by using FCM algorithm integrating with apperceive characteristics human visual system and related local image. According quantization rule paper, information scrambled watermarking adaptively corresponding position. The method finding process extracting same as watermark....