陳永昇

Semi-Supervised Segmentation and Multi-Label Learning with Few-Shot Learning in Medical Image Analysis

資訊科學與工程研究所

半監督醫學影像分割、多標籤醫學影像分類、多標籤少樣本學習、Semi-Supervised Medical Image Segmentation、Multi-Label Medical Image Classification、Multi-Label Few-shot Learning

Gaussian Splatting­-based Stereo Visual SLAM with Keyframe Pose Refinement and Visibility-­guided Gaussian Densification

多媒體工程研究所

雙目視覺同步定位與地圖建構、三維高斯潑濺、關鍵幀位姿優化、特徵重投影一致性、高斯增密、像素感知梯度、stereo visual SLAM、3D Gaussian Splatting、keyframe pose refinement、feature reprojection consistency、Gaussian densification、pixel­-aware gradient

Boundary-Enhanced Semi-Supervised Segmentation via Dual-Decoder Architecture and Data Augmentation for Medical Images

多媒體工程研究所

醫學影像、半監督學習、影像分割、資料擴增、雙解碼器、medical imaging、semi-supervised learning、image segmentation、data augmentation、dual-decoder architecture

A Two­-Stream Decoder with Contour­-Aware Loss for Robust Polyp Segmentation

多媒體工程研究所

大腸鏡檢查、息肉分割、類別不平衡、邊界感知方法、多尺度融合、深度學習、colonoscopy、polyp segmentation、class imbalance、boundary­-aware、multi­-scale fusion、deep learning

Class­-wise Asymmetric Discrepancy Loss for Multi­-label Image Classification under Long­-tailed Distribution

資訊科學與工程研究所

多標籤影像分類、長尾分佈、損失函數、假陽性、胸腔 X 光、multi-label image classification、long-tailed distribution、loss function、false-positive、chest X-ray

Weakly-supervised monocular multimodal depth estimation by recovering metric-scale motion through confidence prediction of iToF depth

多媒體工程研究所

深度估測、飛時測距、跨感知、幾何一致性、時序一致性、絕對尺度、depth estimation、time-of-flight、multi-modality、geometric consistency、temporal consistency、metric scale

SpeechDriven 3D Emotional Facial Animation and Smooth Head Motion

資訊科學與工程研究所

3D 面部動畫、語音驅動的面部動畫、頭部運動、情感解耦、交叉訓練、3D face animation、Speech-driven facial animation、Head motion、emotion disentangle、cross-training

Transforming 2D Bounding Boxes into 3D Representations in Autonomous Driving Scenarios through Geometry Guidance

機器人碩士學位學程

物體檢測、三維資料擴充、自動駕駛、電腦視覺、幾何損失、Object detection、3D data augmentation、Autonomous driving、Computer vision、Geometric loss functions

Dual-­Branch Network with Sampling Strategy for Acute Ischemic Stroke Lesion Segmentation in Full­-Resolution Non-­Contrast Computed Tomography

資訊科學與工程研究所

無顯影劑電腦斷層、急性缺血性中風、雙解析度分支卷積神經網路、採樣策略、深度學習、non­contrast computed tomography、acute ischemic stroke、dual resolution branch network、sampling strategy、deep learning

Music Information Analysis using Deep Neural Networks

資訊科學與工程研究所・共同指導：李建興

深度學習、音樂資訊分析、音樂曲風分類、音樂情緒辨識、音樂節奏估計、音樂節奏生成、Deep learning、Music information analysis、Music genre classification、Music emotion recognition、Music tempo estimation、Music rhythm generation

Accuracy Improvement for EEG Cross-Subject Emotion Classification via Latent Disentangling

資訊科學與工程研究所

腦電波、情緒分類、深度學習、潛在特徵去交錯化、k 折交叉驗證、electroencephalography、emotion recognition、deep learning、latent disentangling、k-fold crossvalidation

Facial Emotion Recognition Based on a Global and Local Attention Network

機器人碩士學位學程

人臉情緒分析、注意力機制、集成網路架構、Facial emotion recognition、Attention、Ensemble network

Triplet Variational Autoencoder with SemanticGuided Features Retrieval for Enhanced Unsupervised Anomaly Detection in Brain MRI

多媒體工程研究所

異常檢測、無監督學習、變分自編碼器、度量學習、交叉注意、核磁共振 成像、Anomaly detection、Unsupervised learning、Variational Autoencoder、Metric ii learning、Cross­attention、MRI

Enhancing Visual Detection with Mixed Stage Partial Network, Parallel FPN, and Background Learning

資訊科學與工程研究所・共同指導：謝君偉

小物件偵測、特徵金字塔、資料擴增、無人機、智慧運輸系統、small object detection、feature pyramid network、data augmentation、unmanned aerial vehicle、intelligent transportation system

Multimodal Sentiment Analysis on Evoked Dataset with Label Smoothing Method

機器人碩士學位學程

多模態情緒分析、標籤平滑化、誘發資料集、模態衝突、噪聲標籤、Multimodal Sentiment Analysis、Label Smoothing、Evoked Dataset、Modality Conflict、Noisy Label

Attention based Multi-Dimensional Cooperative Network for Stereo Matching and Integration with DROID-SLAM for 3D Scene Reconstruction

資訊科學與工程研究所

深度估測、立體匹配、成本體積、三維場景重建、同時定位與地圖構建、depth estimation、stereo matching、cost volume、3D scene reconstruction、SLAM

Advancing Monocular 3D Face Analysis

國際半導體產業學院・共同指導：文卡特什

面部分析、三維重建、遮擋和噪聲魯棒性、加速推理、著色和性別分類、Facial Analysis、3D Reconstruction、Occlusion and Noise Robustness、Accelerated Inference、Colorization and Gender Classification

Anatomy-Prior Hybrid Network for Acute Ischemic Stroke Lesion Segmentation in Non-Contrast Computed Tomography

多媒體工程研究所

無顯影劑電腦斷層、急性缺血性中風、混合維度架構、左右腦比對、深度學習、non-contrast computed tomography、acute ischemic stroke、hybrid network、interhemispheric comparison、deep learning

Robust training strategies for ASPECTS estimation of acute ischemic stroke in multi-­thickness non-­contrast computed tomography with noisy labels

多媒體工程研究所

無顯影電腦斷層、急性缺血性中風、ASPECTS、深度學習、標籤校正、厚度不一致、non­-contrast computed tomography、acute ischemic stroke、deep learning、label calibration、thick impact adjustment

Weakly Supervised Learning of Cross-Modal Depth Estimation using RGB-D Sensors

工學院機器人碩士學位學程・共同指導：邱維辰

深度估測、飛時測距、跨感知、幾何一致性、知識蒸餾、depth estimation、time-of-flight、cross-modal、geometric consistency、knowledge distillation

Self-supervised learning of joint optical flow estimation and stereo matching using attention based estimator

多媒體工程研究所

多任務學習、聯合學習、非監督式學習、注意力機制、立體匹配、光流估計、joint learning、unsupervised、self-supervised、attention、stereo matching、optical flow

Improve Accuracy of Motor-Imagery EEG Signals Classification with Different Parameters in Neural Networks

資訊科學與工程研究所

腦機介面、運動想像、深度學習、腦電波、卷積神經網路、Brain Computer Interface、Motor-Imagery、Deep Learning、Electroencephalography(EEG)、convolution neural network

APU-Net：Anatomy Prior U-Net for Acute Ischemic Stroke Segmentation using Non-contrast Computed Tomography

多媒體工程研究所

無顯影電腦斷層、急性缺血性中風、醫學影像分割、深度學習、左右腦比對、卷積神經網路、non-contrast computed tomography、acute ischemic stroke、medical image segmentation、deep learning、interhemispheric comparison、convolutional neural network

Learning of Subject-Independent Representation in EEG Emotion Classification

多媒體工程研究所

深度學習、腦電波、情緒分類、跨受試者、領域自適應、deep learning、electroencephalography、emotion classification、subject-independent、domain adaptation

Self-Supervised Stereo Matching on Light-Weight CNN

資訊科學與工程研究所

自監督、立體匹配、即時深度估計、輕量化、Self-supervised、Stereo matching、Real-time、Depth estimation、Light-weight

Autonomous Obstacle Avoidance for Drone Navigation based on Real-time Depth Estimation

工學院機器人碩士學位學程

無人機自主避障、無人機導航、即時深度預估、Autonomous Obstacle Avoidance for Drone、Drone Navigation、Real-time Depth Estimation

StyleGAN-v2 Based Reconstruction of Facial Images from fMRI with Improved Attribute Consistency

生醫工程研究所

功能性磁振造影、視覺刺激、面部特徵、重建人臉影像、深度學習、fMRI、visual stimulation、reconstructed face image、facial attribute、deep learning

Sparse Regularization Multilayer Perceptron with Channel Selection for EEG Emotion Recognition

資訊科學與工程研究所

情緒辨識、腦電波、多層感知器、通道選擇、深度學習、Emotion Recognition、EEG、Multilayer Perceptron、Channel Selection、Deep Learning

Glossokinetic-potential-based Human Computer Interface

生醫工程研究所

舌動訊號、深度學習、人機互動介面、Glossokinetic potential、Deep learning、Human Computer Interface

Deep Learning Model for CT Brain Feature Learning and AIS Region Segmentation

生醫工程研究所

深度學習、影像特徵擷取、影像切割、急性缺血性中風、Deep Learning、Feature Extraction、Image Segmentation、Acute Ischemic Stroke

Multi-resolution Residual Convolutional Neural Network for Brain Source Localization

資訊科學與工程研究所

腦源定位、深度學習、多解析度卷積網路、brain source localization、deep learning、multi-resolution convolutional neural network

Emergency Vocal Sound Events Detection and Classification based on Convolutional Recurrent Neural Network and Transfer learning

多媒體工程研究所

聲音事件辨識、轉移學習、卷積遞歸神經網路、深度學習、Sound Event Detection、Transfer Learning、Convolutional Recurrent neural network (CRNN)、Deep Learning

Automated Detection for Acute Ischemic Stroke using Non-Contrast Computed Tomography and Deep Learning

生醫工程研究所・共同指導：陳稷康

深度學習、急性缺血性中風、Deep learning、ASPECTS、Acute Ischemic Stroke

Efficient Deep Neural Network for Dense Depth Estimation and 3D Model Reconstruction

資訊科學與工程研究所・共同指導：陳冠文

單目場景深度預測、多視角立體重建、三維重建、深度學習、Monocular Depth Estimation、Multi-View Stereo、3D Model Reconstruction、Deep Learning

Deciphering Complex Brain Networks with Magnetic Resonance Imaging, Graph-theoretical Analysis and Deep Learning on Graphs: Applications in Cognitive and Clinical Neuroscience

資訊科學與工程研究所・共同指導：郭立威

大腦網路分析、磁振造影、圖論分析、圖深度學習、brain network analysis、Magnetic Resonance Imaging、Graph-theoretical Analysis、Deep Learning on Graphs

Adversarial-based deep transfer learning for EEG emotion classification

生醫工程研究所

深度學習、遷移學習、情緒分類、腦電波、deep learning、transfer learning、emotion recognition、electroencephalography

Face Image Reconstruction with Consistent Facial Attributes from Functional Magnetic Resonance Imaging

生醫工程研究所

功能性磁振造影、視覺影像重建、重建具一致面部特徵之人臉影像、深度神經網路、Functional Magnetic Resonance Imaging、Visual Reconstruction、Reconstruction with Consistent Facial Attributes、Deep Neural Network

Automatic Camera Calibration Using Flat Panel Displays

資訊科學與工程研究所・共同指導：莊仁輝

相機校正、魚眼相機、雙眼相機、Camera Calibration、Fisheye Camera、Stereo Camera

Deep Neural Network with Attention Mechansim and LSTM for Temporal Information Exploration in Classification of Motor-Imagery EEG

資訊科學與工程研究所

腦機介面、運動想像、深度學習、捲積式神經網絡、遞迴式神經網絡、長短期記憶模型、注意力機制、模型解釋、特徵擦拭、Brain-computer interface、Motor-imagery、Deep learning、Convolutional neural network、Recurrent neural network、Long-short term memory、Model interpretation、Representation erasure

FAS2D: Fast and Accurate Sparse-to-dense Depth Estimation

多媒體工程研究所

深度學習、深度估測、稀疏深度、單目視覺、Deep Learning、Depth Estimation、Sparse Depth Samples、Monocular

Learning 3D Geometry for Monocular Depth Estimation

資訊科學與工程研究所

深度學習、深度預測、稀疏深度資訊、學習幾何資訊、Deep Learning、Depth Estimation、Sparse depth information、Learning Geometry Constraint

Effective Deep Network Architectures for Object Recognition and Stereo Depth Estimation

資訊科學與工程研究所

深度卷積網路、物體辨識、立體深度估算、深度學習、deep convolutional neural network、object recognition、stereo depth estimation、deep learning

Model Reconstruction Using Visual SLAM and Deep Stereo Matching

資訊科學與工程研究所

環境模型重建、同步定位與地圖構建、立體匹配、深度神經網路、Model Reconstruction、Visual SLAM、Stereo Matching、Deep Neural Network

Deep Neural Network for Visual Reconstruction from fMRI Data

生醫工程研究所

功能性磁振造影、視覺影像重建、深度神經網路、fMRI、Visual Reconstruction、Deep Neural Network

Domain Transfer for Semantic Segmentation in Aerial Images Using Drone Simulator

多媒體工程研究所

電腦視覺、語義分割、無人飛行機、領域遷移、Computer vision、semantic segmentation、drone、domain transfer

Percussion Rhythm Reconstruction from MEG Data Using Auxiliary Classifier GAN

生醫工程研究所

腦磁波、節奏重建、生成式對抗網路、MEG、Rhythm Reconstruction、GANs、Generative Adversarial Networks

Large-scale Surveillance System of Environment with Reconstructed 3D Model

多媒體工程研究所

三維場景建模、運動恢復結構、監控系統、3D reconstruction、structure from motion、surveillance system

Perspective Correction for 3D Pedestrian Surveillance System

資訊科學與工程研究所

監控系統、三維模型、行人偵測、Surveillance System、3D Model、Pedestrian Detection

VR System for 3D Surveillance

資訊學院資訊學程

虛擬實境、監控系統、三維、VR、surveillance system、3D

Innovative Camera Calibration Based on Projective Geometry

資訊科學與工程研究所・共同指導：莊仁輝、林建州

相機校正、魚眼相機、墨線儀、camera calibration、fisheye camera、laser level

Autoencoder-based Brain Source Localization Using Magnetoencephalographic Data

生醫工程研究所

腦磁波、訊號源定位、自動編碼機、MEG、Source Localization、Autoencoder

Multivariate Pattern Analysis of fMRI Data Using Deep Neural Network

生醫工程研究所

深度神經網路、卷積神經網路、功能性磁振造影、多變量模式分析、fMRI、deep neural network、CNN、MVPA

Brain-Computer Interface System Based on Multi-class Motor Imagery Classification of Brain Source Activity Using Maximum Contrast Beamformer

生醫工程研究所

腦機介面、腦電波、動作想像、BCI、EEG、motor-imagery

Object Model Reconstruction System with High-resolution Texture Mapping using RGB-D Sensor

資訊科學與工程研究所

模型重建、紋理貼圖、多視點對位、全域最佳化、3D Model Reconstruction、Texture Mapping、Multi-view lignment、ICP、Global Optimization

Computer Vision Techniques for Detection and Recognition of Drinking Activity

電機資訊國際學程

電腦視覺、影像處理、辨認、偵測、深度資訊、Computer Vision、Image Processing、object detection、object recognition、depth information

An Augmented Reality-based System for Art-exhibition Guidance by Computer Vision Techniques Using Smart Glasses

多媒體工程研究所・共同指導：蔡文祥

電腦視覺、室內導覽、擴增實境、智慧眼鏡、Computer Vision Techniques、Indoor Guidance、Augmented Reality、Smart Glasses

Fast Construction of Smooth 3D Whole Human-body Color Models by A Two-level Multi-Kinect System

多媒體工程研究所・共同指導：蔡文祥

多Kinect系統、人體模型建構、顏色過濾距離權重相關係數、邊緣偵測、膚色偵測、三維掃描、三維列印、Multi-Kinect system、Human-body model construction、Color-filtered distance-weighted correlation、Edge detection、Skin-color detection、3D scanning、3D printing

Manifold Encoding and Decoding for Investigation into Information Processing in Human Brain using MEG

資訊科學與工程研究所・共同指導：陳麗芬

大腦、腦磁圖、流形、編碼、解碼、視覺處理、人臉處理、brain、MEG、manifold、encoding、decoding、visual processing、face processing

Automatic Program Director System in Virtual Studio using RGB-D Sensor

資訊科學與工程研究所

虛擬攝影棚、彩色深度感測器、去背、手勢辨識、virtual studio、Kinect、background removal、hand gesture recognition

Vehicle Around View System with Dynamic 3D Projection Model

資訊科學與工程研究所

環車視訊、形變、全景圖、立體匹配、Around view、Deformation、Panorama、Stereo matching

The effects of age on visual-spatial working memory (VSWM)：A cross-sectional fMRI study

生醫工程研究所・共同指導：黃植懋

工作記憶、老化、功能性磁振造影、多尺度熵、working memory、aging、fMRI、Multi-scale-entropy

Spatiotemporal Imaging of Functional Connectivity Using Electromagnetic Brain Signals

資訊科學與工程研究所・共同指導：陳麗芬

腦磁圖、活動源造影、功能性連結、神經網路、光束構成法、獨立成份分析、Magnetoencephalography、Source imaging、Functional connectivity、Neural network、Beamformer、Independent component analysis

Moving Object Detection Using Optical Flow Model Captured by a Moving Camera

多媒體工程研究所

移動物體、光流法、Moving Object、Optical Flow

Objective Assessment of Menstrual Pain Scale by Resting-state Brain Activity Estimated from MEG

資訊科學與工程研究所

腦磁波、疼痛評估、經痛、休息狀態、樣本熵、楔前葉、額上回、預設模式網絡、多元線性回歸、MEG、Pain assessment、Menstrual pain、Resting state、Sample entropy、Precuneus、superior frontal gyrus、Default mode network、Multiple linear regression

Fast Diffusion Spectrum MRI Technology using Dictionary-based Compressive Sensing

生醫工程研究所・共同指導：郭立威

磁振造影、擴散頻譜磁振造影、壓縮感測、字典式基底、擴散機率分佈、方向機率分佈、magnetic resonance imaging、diffusion spectrum MRI、compressive sensing、dictionary-based、probability density function、orientation density function

Camera Calibration and Image Stitching of Vehicle Surrounding Monitoring System

資訊科學與工程研究所

監控系統、魚眼攝影機校正、影像接合、Surrounding monitoring system、Fisheye camera calibration、Image stitching

Automatic Co-registration of EEG-MRI data using RGB-D Camera

資訊科學與工程研究所

色彩與深度感測器、磁振造影、腦磁波圖、對位、RGB-D camera、MRI、EEG、Co-registration

3D Object Model Reconstruction with Texture Mapping using RGB-D Camera

資訊科學與工程研究所

模型重建、reconstruction、kinect

EEG-based Person Identification System and Its Longitudinal Adaptation

生醫工程研究所

腦電波、身分辨識、EEG、Person identification

Objective Assessment of Menstrual Pain Scale

生醫工程研究所

靜息態、經痛、腦磁波、多尺度熵、resting、menstrual pain、MEG、multi-scale entropy

3-D Object Model Reconstruction and Immersive Interaction using Kinect Sensor

資訊科學與工程研究所

對位、整合、模型重建、透視投影、互動、registration、integration、model reconstruction、perspective projection、interaction

A Fast and Accurate Algorithm for Diffeomorphic and Symmetric Non-rigid Registration of Brain Magnetic Resonance Images

資訊科學與工程研究所

對位、微分同構、磁振造影、registration、diffeomorphism、magnetic resonance image

Construction of Brain Templates using T1-weighted and DT MRI Data

生醫工程研究所

腦影像模板、座標空間、T1權重、擴散張量磁振造影、磁振造影、客製化、非剛性對位演算法、微分同構、對稱性、對數歐幾里德架構、非正特徵值、正定性、正規化、brain templates、stereotactic space、T1-weighted、DTI、MRI、customized、non-rigid registration algorithm、diffeomorphic、symmetric、Log-Euclidean framework、non-positive eigenvalues、positive definiteness、normalization

Video Content Recovery and Modification by Object Inpainting

資訊科學與工程研究所・共同指導：廖弘源

影片修補、video inpainting

Emotion Recognition using Magnetoencephalographic Recordings Evoked by Affective Visual Stimuli

生醫工程研究所

情緒辨識、腦磁波、事件相關磁場、自迴歸模型、頻譜能量密度、主成分分析、局部線性嵌入、Emotion Recognition using、Magnetoencephalographic、event-related magnetic fields、autoregressive (AR) model、power spectral density、principal component analysis、linear embedding

Person Identification using Electroencephalographic Signals Evoked by Visual Stimuli

生醫工程研究所

腦電圖、身份辨識、視覺刺激誘發電位、EEG、person identification、VEP

Detection of Cyclic Alternating Pattern in Electroencephalographic Signals

資訊科學與工程研究所

睡眠腦電圖、微觀分析、循環交替模式、Sleep EEG、Microstructure Analysis、Cyclic Alternating Pattern

Construction of Customized Brain Template from Magnetic Resonance Images

資訊科學與工程研究所

腦模板、磁振造影、客製化、brain template、MRI、customized

Long-term Object Tracking in Video with Unsteady Frame Rate

資訊科學與工程研究所

長程追蹤、object tracking、mean shift、unsteady frame rate、unstable frame rate