A5082090342- Human Pose and Action Recognition
- Advanced Neural Network Applications
- Cryptography and Data Security
- Context-Aware Activity Recognition Systems
- Complexity and Algorithms in Graphs
- Video Surveillance and Tracking Methods
- Ferroelectric and Negative Capacitance Devices
- Natural Language Processing Techniques
- Inertial Sensor and Navigation
- Cryptographic Implementations and Security
- Optical measurement and interference techniques
- Speech Recognition and Synthesis
- Water Systems and Optimization
- Advanced Memory and Neural Computing
- Maritime Navigation and Safety
- Chaos-based Image/Signal Encryption
- Flow Measurement and Analysis
- Cryptography and Residue Arithmetic
- Image Enhancement Techniques
- Anomaly Detection Techniques and Applications
- Non-Invasive Vital Sign Monitoring
- Topic Modeling
Southern University of Science and Technology
2021-2025
Shandong University of Traditional Chinese Medicine
2024
University of Hong Kong
2022-2024
China Jiliang University
2023
Various industrial and domestic applications call for optimized lightweight video LSTM network models on edge. The recent tensor-train method can transform space-time features into tensors, which be further decomposed low-rank analysis rank selection of tensor is however manually performed with no optimization. This paper formulates a search algorithm to automatically decide ranks consideration the trade-off between accuracy complexity. A fast method, called RankSearch, developed find...
Tripping or falling is among the top threats in elderly healthcare, and development of automatic fall detection systems are considerable importance. With fast Internet Things (IoT), camera vision-based solutions have drawn much attention recent years. The traditional video analysis on cloud has significant communication overhead. This work introduces a lightweight network based spatio-temporal joint-point model to overcome these hurdles. Instead detecting motion by Convolutional Neural...
Falling is ranked highly among the threats in elderly healthcare, which promotes development of automatic fall detection systems with extensive concern. With fast Internet Things (IoT) and Artificial Intelligence (AI), camera vision-based solutions have drawn much attention for single-frame prediction video understanding on by using Convolutional Neural Network (CNN) 3D-CNN, respectively. However, these methods hardly supervise intermediate features good accurate efficient performance edge...
Large Language Models (LLMs) have greatly advanced the natural language processing paradigm.However, high computational load and huge model sizes pose a grand challenge for deployment on edge devices.To this end, we propose APTQ (Attention-aware Post-Training Mixed-Precision Quantization) LLMs, which considers not only second-order information of each layer's weights, but also, first time, nonlinear effect attention outputs entire model.We leverage Hessian trace as sensitivity metric...
Traditional neural networks deployed on CPU/GPU architectures have achieved impressive results various AI tasks. However, the growing model sizes and intensive computation presented stringent challenges for deployment edge devices with restrictive compute storage resources. This paper proposes a one-shot training-evaluation framework to solve architecture search (NAS) problem in-memory computing, targeting emerging resistive random-access memory (RRAM) analog platform. We test inference...
Ensuring the accuracy of flow measurement is crucial to promoting high-quality cigarette production. In order monitor working status flowmeters, this paper proposes an anomaly detection method based on sliding-window recursive Lasso (Least absolute shrinkage and selection operator), which able track changes in flowmeter operating conditions by self-adapting model parameters observed measurements. Due frequent mode switch high sampling frequency data, introduces strategy remove effect...