李慶鴻

Cluster-Based Adaptive Image Captioning for Enhancing Visual Content Descriptions in Multimodal RAG

電控工程研究所

檢索增強生成、多模態文件問答、自適應影像描述、影像聚類、提示工程、Retrieval-Augmented Generation、Multimodal Document Question Answering、Adaptive Image Captioning、Image Clustering、Prompt Engineering

Early Yield Risk Prediction in Semiconductor Manufacturing Using WAT Data and Hybrid Machine Learning Approaches

電機學院碩士在職專班

半導體、機器學習、良率預測、不平衡資料集、晶圓測試、特徵篩選、Semiconductor、Machine Learning、Yield Prediction、Imbalanced Dataset、Wafer Acceptance Test (WAT)、Final Testing (FT)、Feature Selection

Adaptive Vibration Suppression Framework for Wafer-Handling Robots

電控工程研究所

晶圓傳輸機械手臂、主動式振動抑制、修正型加速度回授、代理模型、貝氏優化、Wafer-Handling Robot、Active Vibration Control、Modified Acceleration Feedback、Sequential Model-Based Optimization、Surrogate Model、Bayesian Optimization

Path Accuracy Improvement for Serial-Link Robot Manipulators

電控工程研究所

串聯式機械手臂、軌跡精度、量測系統、伺服調機、路徑補償、迭代式補償、Serial-Link Robot Manipulato、Path Accuracy、Measurement System、Servo Tuning、Path Compensation、Iterative Compensation

SSFANet: Selective Spatial Feature Aggregation on U-Net with Attention Modules for Low-Light Image Enhancement

電機學院碩士在職專班

低光源圖像增強、注意力機制、卷積神經網路、編解碼器、選擇性空間特徵聚合模組、密集並行注意力區塊、Low-Light Image Enhancement、Attention Mechanism、Convolutional Neural Network、Encoder-Decoder、Selective Spatial Feature Aggregation Module、Dense Parallel Attention Block

Inspection System Development for Exterior Wall Tile Defect Detection

機器人碩士學位學程

磁磚瑕疵檢測、多感測器檢測系統、聲音訊號分類、無人機、ROS2、Tile defect inspection、Multi-sensor system、Audio Classification、UAV

Scalable Long Time Series Diffusion Models for Vibration Signal Generation

電控工程研究所

時序資料生成、降噪擴散機率模型、類別條件生成、振動訊號生成、刀具磨耗預估、Time Series Generation、Denoising Diffusion Probabilistic Models、Class-Conditioned Generation、Vibration Signal Generation、Tool Wear Estimation

EEG Data Augmentation for Depression Diagnosis Using Generative Adversarial Networks

電控工程研究所

重度憂鬱症、腦電圖、輔助診斷系統、通道選擇、生成對抗網路、資料增強、Major Depressive Disorder、Electroencephalography、Computer-Aided Diagnosis、Channel Selection、Generative Adversarial Network、Data Augmentation

Semi-Supervised Domain Adaptation for Cross-Dataset EEG-based Cognitive Impairment Recognition

電控工程研究所

腦電圖、失智症、頻域分析、深度學習、領域自適應、半監督式學習、EEG、Dementia、Frequency Domain Analysis、Deep Learning、Domain Adaptation、Semi-Supervised Learning

Research on Cooking Robots Based on Autonomous Cutting Direction Generation

電機資訊國際碩士學位學程

烹飪機器人、深度強化學習、多步切割、自動編碼器、形狀匹配、Cooking Robot、Deep Reinforcement Learning、Multi-step Cutting、Autoencoder、Shape Matching

Intelligent Diagnostics and Servo Tuning for Enhancing Dynamic Path Accuracy of Six-Axis Robot Manipulators

電控工程研究所

機械手臂、伺服參數優化、動態精度提升、拉格朗日插值法、伺服性能優化、一維卷積神經網路、虛擬伺服系統、Robotic Manipulators、Servo Tuning、Dynamic Path Accuracy、Lagrange Interpolation、Servo Diagnostics、One Dimensional Convolutional Neural Networks、Virtual Servo-drive System

Speed Strategy for Optimizing Endurance Based on the Electrical Power Model of Spraying Drones

電控工程研究所

植物保護無人機、續航力、電功率模型、速度策略、最佳飛速與噴率、Spraying drone、Endurance、Electrical power model、Speed strategy、Optimal flight speed and spray rate

Development of an Autonomous Rail Tracking and Foreign Object Detection System Based on Quadcopter Integrated with Edge Devices

電控工程研究所

鐵路安全、四軸無人機、自主飛行、異物入侵檢測、即時預警、railway safety、quadrotor aircraft、autonomous flight、foreign object detection、real-time alerts

Indoor Patrol Robot with Face Recognition for Low-light Condition

電控工程研究所

低光源環境、人臉識別、影像增強、生成對抗網絡、室內巡邏系統、Low-light environments、Face recognition、Image enhancement、Generative Adversarial Networks、Indoor patrol system

Point Cloud Semantic Segmentation Map for Indoor Robots

機器人碩士學位學程

圖像語義分割、視覺同步定位與建圖、定位與路徑 規劃、點雲語義分割地圖、Image Semantic Segmentation、Simultaneous Localization and Mapping、Localization and Path Planning、Point Cloud Semantic Segmentation Map

A Closed-Loop Rehabilitation System Based on SSVEP Brain-Computer Interface

電控工程研究所

腦電波、上肢復健、特徵篩選、閉迴路系統、穩態視覺誘發電位、事件相關非同步、Brainwaves、Upper limb rehabilitation、Feature selection、Closed-loop system、Steady-state visual evoked potentials、Event-related desynchronization

Applications of Federated Learning Edge Computing in Tool Wear Estimation System

電控工程研究所

刀具磨耗、邊緣運算、聯邦學習、領域適應、知識蒸餾、Tool Wear、Edge Computing、Federated Learning、Domain Adaptation、Knowledge Distillation

Manufacturing Parameters Optimization in Surface Mount Technology Using Surrogate-Based Optimization

電控工程研究所

表面貼裝技術、錫膏印刷、回流焊接、代理模型、粒子群演算法、類神經網路、Surface mount technology、Solder paste printing、Reflow soldering、Surrogate model、Particle swarm optimization、Neural network

Depth of Anesthesia Estimation Using Band Power Ratios and Unsupervised Learning-guided Classifier

電控工程研究所

麻醉深度監測、腦電圖、患者狀態指數、密度頻譜陣列、個體化差異、模糊C-均值算法、XGBoost 分類器、Depth of Anesthesia Monitoring、Electroencephalography、Patient State Index、Density Spectral Array、Individual Difference、Fuzzy C-means、XGBoost Classifier

Binary One-Dimensional Convolutional Neural Network Accelerator for Bearing Fault Signal Diagnosis

電控工程研究所

模型壓縮、二值化神經網路、軸承故障診斷、FPGA、Model Compression、Binary Neural Networks、Bearing Fault Diagnosis

Generalized Optimal EEG Channels for Motor Imagery Brain-Computer Interface by Using Machine Learning Techniques

電控工程研究所

腦電圖、腦電圖通道選擇、廣義最佳腦電圖通道、運動心像、偏側化準備電位、NSGA-II 演算法、Electroencephalography (EEG)、Channel selection、Generalized optimal EEG channels、Motor Imagery (MI)、Lateralized Readiness Potential (LRP)、EEG-Net、NSGA-II Algorithm

Integrating Deep Learning and Statistical Analysis for Structural Damage Identification Approach

電控工程研究所

結構損傷識別、振動訊號、訊號差、感測器可辨識度、卷積神經網路、structural damage identification、vibration signals、signal dierence、sensor discriminability、convolutional neural networks

Collision Detection Technique Development of Robot Manipulator Using Deep Learning

電機學院電機與控制學程

機械手臂、碰撞偵測、動力學模型、卷積類神經網路、工具重力補償、粒子群演算法、robot manipulator、collision detection、robot dynamic model、convolution neural network、tool gravity compensation、particle swarm optimization

An Online Domain Adaptation Approach for Machine Health Monitoring with Incomplete Target Domain Data

工學院機器人碩士學位學程

遷移學習、領域適應、機台健康監測、刀具磨耗、軸承故障診斷、Transfer Learning、Domain Adaptation、Machine Health Monitoring、Tool Wear、Bearing Fault Diagnosis

A Novel Personalized Federated Learning Algorithm with Privacy Constraint

電控工程研究所

隱私保護、資料異質性、梯度平均、個性化聯邦學習、通訊成本、Privacy-Preserving、Heterogeneity、Gradient Averaging、Personalized Federated Learning、Communication Cost

A Novel Label Smoothing Technique for Machine State Degradation

電控工程研究所

機台老化、感測器挑選、平滑標籤、卷積類神經網路、刀具磨耗、軸承健康監測、machine degradation、sensor selection、label smoothing、convolution neural network、tool wear、bearing health monitoring