- Advancements in Battery Materials
- Multimodal Machine Learning Applications
- Advanced Memory and Neural Computing
- Advanced Neural Network Applications
- Topic Modeling
- Natural Language Processing Techniques
- Image and Object Detection Techniques
- Neural Networks and Applications
- Extraction and Separation Processes
- Advanced Battery Technologies Research
- EEG and Brain-Computer Interfaces
- Recycling and Waste Management Techniques
- Advanced Measurement and Detection Methods
- Visual Attention and Saliency Detection
- Emotion and Mood Recognition
- Fire Detection and Safety Systems
- Industrial Vision Systems and Defect Detection
- Analog and Mixed-Signal Circuit Design
- Advanced Decision-Making Techniques
- Remote Sensing and LiDAR Applications
- Smart Grid and Power Systems
- Cognitive Computing and Networks
- IoT-based Smart Home Systems
- Robotics and Sensor-Based Localization
- Electric Vehicles and Infrastructure
Hubei University
2024-2025
Zhejiang University
2017-2024
Hubei University of Technology
2024
Wuhan University of Technology
2020-2022
Changzhou Institute of Technology
2017
Yunnan Normal University
2017
Mental health problems are an increasingly common social issue severely affecting and well-being. Multimedia processing technologies via facial expression show appealing prospects in the consumer field for mental monitoring, while still suffer from intensive computation low energy efficiency. This paper proposes energy-efficiency memristive sequencer network (EMSN) human emotion classification, which offers environmentally friendly approach consumers with cost easily deployable hardware....
Video sentiment analysis can effectively establish the relationship between emotion state and multimodal information, while still suffer from intensive computation low efficiency, due to von Neumann computing architecture. Here, we present a brain-inspired hierarchical interactive in-memory (IMC) system, which efficiently solve 'von bottleneck', enabling cross-modal interactions semantic gap elimination. First, 1T1M synapse array is fabricated using cost-effective, highly stable, flexible,...
Terraces are the major land-use type of agriculture and support main agricultural production in southeast southwest China. However, due to smallholder farming, complex terrains, natural disasters illegal land occupations, a light-weight low cost dynamic monitoring terraces has become serious concern for systems above area. In this work, we propose small unmanned aerial vehicle (UAV) based multi-temporal image registration method that plays an important role transforming images into one...
Fluorine-doped carbon-coated LiFePO 4 materials are regenerated by sintering residual PVDF after separating spent electrode strips via methanol-citric acid, with a good capacity of 141.5 mA h g −1 at 1C and retention rate 99.6% 100 cycles.
Development of simple hydrometallurgy strategies for recovering cathode materials in spent LiFePO4 (LFP) batteries is highly desired, but the removal impurity metal components remains a complex and challenging task. Herein, we proposed passivation-driven mechanism high-selective leaching an Al(OH)3 template self-dissolution hydrothermal method regenerating hollow-structure LFP materials. Experimentally, 99.0% Li 98.7% Fe can be leached out while only 4.0% Al from electrode sheets under...
Abstract Due to the highly complex and non‐linear physical dynamics of lithium‐ion batteries, it is unfeasible measure state charge (SOC) directly. Designing systems capable accurate SOC estimation has become a key technology for battery management (BMS). Existing mainstream approaches still suffer from limitations low efficiency high‐power consumption, owing great number samples required training. To address these gaps, this paper proposes memristor‐based denoising autoencoder gated...
In recent years, the application of machine translation has become more and widely. Currently, neural multimodal models have made attractive progress, which combines images into deep learning networks, such as Transformer RNN. When considering in models, they directly apply gate structure or image attention to introduce feature enhance effect. We argue that it may mismatch text features since are different semantic space. this paper, we propose a coordinated representation enhanced approach...
The traditional inspection method of substation equipment mainly relies on human resources, which has shortages as high work intensity, low efficiency and lack real-time. In this paper, the framework whole area intelligent system is designed, combines real-time performance fixed camera with mobility robot, complements each other's advantages. On basis, through combination improved deep learning Fast-RCNN algorithm effective part algorithm, recognition typical appearance defects realized....
Transformer is one of the most important power equipment in substation. Generally, workers substation are supposed to inspect transformer regularly and try avoid abnormal operation equipment, including state inspection infrared detection. This contributes waste manpower task cannot be maintained frequently. paper has proposed a defects detection method based on visible fusion images. The consists deep learning model which focuses It able detect overheating automatically using binocular...
With the increasing availability of images, multimodal machine translation (MMT) is leading a vibrant field. Model structure and information introduction are hotspot focused by MMT researchers nowadays. Among existing models, transformer model has reached state-of-the-art performance in many tasks. However, we observe that based on highly unstable since sensitive to fluctuation hyper-parameters especially number layers, dimension word embeddings hidden states, multi-heads. Moreover,...
Electric vehicles have been developing rapidly in recent years, the State-of-Charge (SOC) of lithium-ion batteries can related to safety and reliability electric vehicles. With development field machine learning, data-driven methods often use learning platforms obtain relationship between battery measurement signals SOC from them achieve accurate estimation SOC. To further explore potential estimation, an improved model using Nesterov Accelerated Gradient (NAG) algorithm based Bidirectional...
The paper has proposed an on-line temperature monitoring device, which can be used to acquire thermal information. Specifically, this device is a binocular IR camera, whose size only 3 cm×3 cm×4 cm. powered by remote Power over Ethernet (PoE), installed near measured objects and responsible for collecting both visible-light images array data of objects. Additionally, the transmitted through same cable, convenient. Furthermore, IR-fusion module in able fuse image collected camera produces...
Automotive ICs work in wide ambient temperature range up to 150[Formula: see text]C. It is important design an over protection mechanism for the reliability of and systems. A thermal module automotive reported this paper. Dual channel detection decision scheme was designed based on band gap voltage reference. Precision point set by serial resistors variations power supply, process were removed resistor ratio. The implemented CSMC 0.5 [Formula: text] 60 V BCD process, incorporated a CAN...
This paper has put forward a method of switch state identification, in which the proposed conjecture is based on generalized Hough transform and decision forest. In this method, position object determined by each other's can achieve purpose checking accurately. The experimental results have illustrated that high recognition rate, good robustness generalization.
This paper has proposed a real-time detection method for the handling mode of surge arrester. Surge arrester is one most important elements in power system, which always used to protect electrical equipment from high transient overvoltage and limit continuation time. However, because its special structure, lead line easy be damaged process construction. Under this circumstance, it necessary design strict automatic system. Io order satisfy these requirements, combines YOLOv3 object algorithm...