劉志尉

MambaE2VID: Selective State Space Model for High-Speed, High-Quality Event-to-Video Reconstruction

電子研究所

事件相機、選擇性狀態空間、高動態範圍、影像重建、Event Camera、High Dynamic Range(HDR)、Selective State Space Model、Double Attention Modules(DAMs)、Image Reconstruction

Mamba-based Spatio-Temporal Modeling for Event Camera

電子研究所

事件相機、物件偵測、Mamba、狀態空間模型、時空建模、Event-based Camera、Object Detection、State Space Model、Spatio-temporal Modeling

Combining Asynchronous GNN, Temporal CNN, and Mamba for Event-Based Object Detection

電子研究所

事件相機、物件辨識、深度學習、圖神經網路、狀態空間模型、Event Camera、Object Detection、Deep Learning、Graph Neural Network、State Space Models

A Sequence-Based High-Rate Hierarchical QC-LDPC Code Construction for NAND Flash Memory

電子研究所・共同指導：余俊宏

階層式類循環 LDPC 碼、PEG 演算法、函數序列建構、短循環、長碼長、高碼率、快閃記憶體、Hierarchical Quasi-Cyclic LDPC Codes (HQC-LDPC)、Progressive Edge-Growth (PEG) Algorithm、Sequence-Based Construction、Short Cycles、Long Code Length、High Code Rate、NAND Flash Memory

A High-Speed Hardware-Accelerated Transformer for Efficient Linear and Non-Linear Operations in Large Language Models

電子研究所

大型語言模型、硬體加速器、非線性運算加速、Transformer、Softmax、GELU、Layer Normalization

A Construction Method Of Quasi-Cyclic LDPC Codes Based On Sidon Sequence With Girth Eight

電子研究所

準循環低密度奇偶校驗碼、LDPC碼、Sidon序列、短環、Quasi-Cyclic LDPC code、Low-density parity-check code、Sidon sequence、cycle

Sparse Distributed Memory Using Constant-Weight Code for Low Power Spiking Neural Network

電子研究所

稀疏分散式記憶體、M-of-N碼、脈衝神經網絡、非監督式學習、脈衝時序依賴可塑性學習方法、Sparse Distributed Memory、M-of-N code、Spiking Neural Network、Unsupervised Learning、STDP Learning Rule

A Hardware Implementation of Image Signal Processing Neural Network Using RISC-V Custom ISA Extension

電子研究所

人工智慧圖像訊號處理、RISC-V、自訂擴展指令集、硬體加速器、現場可程式化邏輯閘陣列、AI-ISP、Custom ISA extension、Hardware accelerator、FPGAs

High-Quality Image Reconstruction for Event-based Cameras

電子研究所

事件相機、高時間解析度、高動態範圍 (HDR)、影像重建、Event-based camera、High temporal resolution、High dynamic range (HDR)、Image reconstruction

Multi-Object Tracking of Pedestrians with Irregular Motion

人工智慧技術與應用碩士學位學程

多目標追蹤、物件偵測、數據關聯、行人重識別、深度學習、Multiple Object Tracking、Object Detection、Data Association、Person Re-identification、Deep Learning

City-Scale Multi-Camera Multi-Target Tracking

電子研究所

物件偵測、重識別、物件追蹤、跨相機追蹤、電腦視覺、Object detection、Re-identification、Object tracking、Cross-camera tracking、Computer vision

A Study on Interference Cancellation and Object Tracking for FMCW Automotive Radar

電子研究所・共同指導：林大衛

FMCW車用雷達訊號處理、物體追蹤、干擾消除、干擾抑制方法、最小平方法估計、卡爾曼濾波器追蹤、FMCW Automotive Radar Signal Processing、Object Tracking、Interference Cancellation、Interference Mitigation Method、Least Squares Estimation、Kalman Filter Tracking

Robust AES Circuit with Signal Integrity Check for Flash Memory Controller

電子研究所

進階加密標準、可容錯之演算法、錯誤偵測碼、差別錯誤分析、電路設計、綜合性電路、Advanced encryption standard、Error-detection algorithm、Error-detection coding、Differential fault analysis、Architecture design、Comprehensive design

Low-Probability of Intercept Radar Signal Classification with Reinforcement Learning and Deep Learning Algorithm

電子研究所

深度學習、卷積神經網路、強化學習、LPI雷達訊號、人工智慧、deep learning、CNN、reinforcement learning、LPI radar signal、AI

Detection of Nonlinearly Distorted OFDM Signals due to One-Bit ADCs Under Known and Unknown CSI

電子研究所・共同指導：林大衛

一位元類比數位轉換器、類比數位轉換器、最大期望算法、訊號檢測、正交分頻多工、通道狀態資訊、one-bit ADC、ADC、EM algorithm、signal detection、OFDM、CSI

Sparse and Irregular Matrix Multiplication Accelerator

電子研究所

矩陣乘法運算、特殊應用硬體架構、低精確度浮點數運算、乘累加器、多重精確度運算、深度學習計算、Matrix Multiplication、Application-Specific hardware、Low precision Floating-Point Arithmetic、Multiply-Accumulate Unit、Multiple-Precision Arithmetic、Deep Neural Network Computing

A High-Parallel Computing In Resistive Random Access Memory Architecture for CNN Application

電子研究所・共同指導：周世傑

人工智慧、記憶體內運算、電阻式隨機存取記憶體、高平行度、AI、Computing In-Memory、RRAM、High parallel

A Study on Interference Mitigation Radar Systems

電子研究所

抗干擾雷達系統、FMCW雷達、EHC-Frank碼、分數傅立葉轉換、Interference Mitigation Radar Systems、FMCW Radar、EHC-Frank code、Fractional Fourier Transform

Defense Against Adversarial Attacks using FFT

電子研究所

快速傅立葉轉換、對抗性樣本、捲積神經網路、二元遮罩、頻率分析、Fast Fourier Transform、Adversarial Example、Convolutional Neural Network、Binary Mask、Frequency Analysis

An Efficient Frequency Domain Training Framework for Robust Convolutional Neural Networks

電子研究所

卷積神經網絡、頻域、強健性、Convolutional Neural Networks、Frequency Domain、Robustness

Automated LPI Radar Detection and Classification using Fractional Fourier Transform

電子研究所

低截獲率雷達、雷達波型辨識、分數傅立葉轉換、深度學習模型、Low Probability of Intercept (LPI) Radar、Radar Waveform Classification、Fractional Fourier Transform、Deep Learning Model

A Low-complexity Pseudo Direct Learning Digital Pre-distortion Architecture for Nonlinearity and Memory Effect of Power Amplifier in Millimeter Wave Baseband Transmitter

電子研究所・共同指導：周世傑

毫米波、功率放大器、基頻、數位預補償、低複雜度、擬直接學習、硬體實踐、millimeter wave、power amplifier、baseband、digital pre-distortion、low-complexity、pseudo direct learning、hardware implement、802.11ad

Design of mmWave Hybrid Precoder and Combiner for SU-MIMO Systems

電子研究所・共同指導：周世傑

毫米波、混合波束成形、單一使用者多輸入多輸出、mmWave、Hybrid Beamforming、SU-MIMO、802.11ay

Architecture of Jointly Compensation for Analog Front-End Non-Ideal Effects in mm-Wave Single Carrier Baseband Receiver Using Machine Learning Methods

電子研究所・共同指導：周世傑

通訊、不理想效應、支援向量機、5G、Communication、SVM、non-ideal effects

Interference Mitigation for Multi-User Radar System Using EHC-Frank Code with 1.25 GHz Bandwidth

電子研究所

雷達相互干擾抑制、多相序列、相關函數器設計、radar mutual interference mitigation、polyphase sequence、Frank code、Extended Hyperbolic Congruential code、correlator design

Automated RTL Generator for High-Throughput and Area-Efficient Ultra-Long FFT Processor

電子研究所

快速傅立葉變換、高吞吐率、RTL產生器、64K點、多路徑延遲交換、FFT、high throughput、RTL generator、64K-point、MDC

A Low Latency NN-based Cyclic Jacobi EVD Processor for DOA Estimation in Radar System

電子研究所

角度偵測器、特偵值分解、神經網路模型、靜態浮點數運算、DOA Estimation、eigenvalue decomposition、neural network model、static floating point

Design and Implementation of an Efficient Radar Signal Detector Based on Data Variability

電子研究所

雷達訊號偵測器、Radar Signal Detector

The Study of ANSI-Frequency Cepstral Coefficients (AFCCs) and Its Applications

電子研究所

頻率倒譜係數、語者辨識、語音辨識、cepstral coefficients、speaker identification、speech recognition

Energy Efficient CNN Inference Accelerator using Fast Fourier Transform

電子研究所

卷積神經網路、機器學習、加速器、快速傅立葉轉換、Convolutional Neural Network、Machine Learning、Accelerator、Fast Fourier Transform、CNN Inference

High-Throughput 64K-point FFT Processor for THz Imaging Radar System

電子研究所

快速傅立葉轉換、高吞吐率、64K點、FFT、high throughput、64K-point

Phase Noise-Robust Synchronization and Phase Noise Compensation Design for 10Gbps Single Carrier Baseband Receiver at 60 GHz Band

電子研究所・共同指導：周世傑

毫米波、基頻、接收機、相雜訊、同步、單一載波、millimeter wave、baseband、receiver、phase noise、synchronization、single carrier

FFT-Based Multirate Signal Processing for Low Complexity 18-Band Quasi-ANSI S1.11 1/3-Octave Filter Bank

電子研究所

多率濾波器組、助聽器濾波器組、實數訊號分頻處理、助聽器分頻系統、Quasi-ANSI filter bank、Multirate Signal Processin、Low Complexity Quasi-ANSI filter、Analysis filter bank for hearing aids

A High-accuracy and Cost-effective SFP SUMMA Array Processor for CNN Inference Application

電子研究所

神經網路、加速器、靜態浮點數運算、陣列處理器、卷積神經網路、convolution neural network、accerlerator、static floating point、CNN inference、array processor

Low Latency Noise Reduction Algorithm and Its Implementation for Binaural Hearing Aids

電子研究所

雙耳助聽器、降噪演算法、低延遲、雙耳線索、硬體實作、binaural hearing aids、noise reduction algorithm、low latency、binaural cues、hardware implementation

A Study of Programmable Non-Uniform Analysis FIR Filter Bank Using Re-Sampling Scheme

電子研究所

取樣率轉換器、奈奎斯特取樣率、濾波器組、數位助聽器、重新取樣技術、sampling rate conversion、Nyquist rate、filter bank、hearing-aids、re-sampling

Fast Monoscopic 2D Image Depth Estimation Method Based on Edge Defocus Cues

生醫工程研究所

散焦線索、深度估測、單相機、Defocus cue、Depth Estimation、Monoscopic

The Analysis and Design of Time-frequency Decomposition, Noise Reduction, and Adaptive Feedback Cancellation for Hearing-aids

電子研究所・共同指導：周世傑

數位助聽器、助聽器系統、多率濾波器組設計、低複雜度語音降躁、低複雜度回授消除、Digital hearing-aids、Hearing-aid system、Multirate filter bank design、Low-complexity noise reduction、Low-complexity acoustic feedback cancellation

Design and Implementation of Low-Complexity Noise Reduction Algorithm Based on Low-Delay Quasi-ANSI Filter Bank for Binaural Hearing Aids

電子工程學系 電子研究所

降噪、濾波器組、雙耳助聽器、noise reduction、quasi-ANSI filter band、binaural hearing aids

In-Situ Error Detection and Correction for Ultra-Low-Voltage VLSI Design

電子工程學系 電子研究所

極低電壓、現場錯誤偵測、可變延遲設計、Ultra-low-voltage、In-situ error detection、Variable-latency design

FFT based Oscillation detector and Feedback Cancellation with Adjustable Step Size and Probe Injection Method for Digital Hearing Aid System

電子工程學系 電子研究所・共同指導：周世傑

助聽器、回授消除、震盪偵測、尖銳音消除、傅立葉轉換、音高式、Digital Hearing Aid、Feedback Cancellation、Oscillation detector、Howling Suppression、FFT、Pitch Based

Phase Noise Cancellation with Sampling Clock Offset Tracking Equalizer for 60 GHz FBMC System

電子工程學系 電子研究所・共同指導：周世傑

多載波濾波器組、60 GHz、取樣頻率偏移、相位雜訊、FBMC、SCO、Phase Noise

Fast Algorithm and Architecture Design of Motion Estimation for HEVC

電子工程學系 電子研究所

移動估測、高效率視訊編碼、motion estimation、HEVC

An Area-/Energy- Efficient Adaptive FIR Filter using Distributed Arithmetic

電子工程學系 電子研究所

分散式算術、適應性濾波器、有限脈衝響應、最小均方演算法、distributed arithmetic、adaptive filter、finite impulse response、least mean square

Low-power digital signal processor for hearing aids

電子工程學系 電子研究所

低功耗、助聽器、可變延遲設計技術、低延遲濾波器組、Low power、hearing aid、Variable latency design technique、low-delay filter bank

An Ultra-fast Response and Low Power Consumption Digital Low Dropout Regulator for High Speed embedded Memory Applications

電機學院電機與控制學程・共同指導：陳科宏

數位式降壓電路、嵌入式記憶體、低功率、Digital LDO、embedded memory、low power

Design of Multi-Stage Computing Engine for DSP Applications

電子工程學系 電子研究所

多級運算引擎、串流介面單元、Multi-stage Computing Engine、Stream Interface Unit

An Area-/Energy- Efficient CSD Shift-and-Add Computing Engine for FIR Filter Bank

電子工程學系 電子研究所

最少冗餘有號數、系數乘法、可規劃、CSD、Coefficient Multiplication、Programmable

A ROM-Free Logarithmic Functional Unit for 3-D Graphics Applications

電子工程學系 電子研究所

三維圖形、對數函式單元、3D Graphics、logarithmic functional unit

Study on Variable-Latency Speculating Booth Multiplier for Multimedia Applications

電子工程學系 電子研究所

布斯乘法器、猜測、可變延遲、錯誤偵測、Booth multiplier、Speculation、Variable-latency、Error detection

A Lightweight Task Controller for Media-Streaming Applications on Heterogeneous Multicore Processor

電子工程學系 電子研究所

多核心排程、multicore scheduling

Design and Analysis of Dynamic Range Detection Array Multiplier for Stochastic Processors

電機學院電子與光電學程

容錯設計、隨機計算、值域偵測、Error tolerant design、Stochastic computation、Range detection

Design and Implementation of Lightweight Processing Element for Digital Hearing Aids

電子工程學系 電子研究所

助聽器、處理器、hearing aids、processing element

Binary-Mask Based Noise Reduction for Hearing Aids with Binaural Cue Preservation

電子工程學系 電子研究所

雜訊消除、雙耳助聽器、聲音定位、Noise Reduction、Binaural Hearing Aid、Sound Localization

Low-Power Auditory Compensation for Digital Hearing Aids

電子工程學系 電子研究所

助聽器、濾波器組、動態範圍壓縮、低功耗、hearing aid、filter bank、dynamic range compression、low power

Complexity-Effective Auditory Compensation with a Controllable Filter for Digital Hearing Aids

電子工程學系 電子研究所

數位助聽器、聽覺補償系統、低功耗、低運算複雜度、digital hearing aids、auditory compensation、low power、low complexity

Design and Implementation of Speculating Functional Units

電子工程學系 電子研究所

猜測、乘法、高效能、算術單元、進位估計、錯誤補償、關鍵路徑、可變延遲、speculating、speculation、multiplication、multiplier、High performance、functional unit、carry estimation、error compensation、critical path、variable latency