高等數位訊號處理

Advanced Digital Signal Processing

高等數位訊號處理

授課者：丁建均

上課時間：星期三下午 3:30~6:20 (3:30~4:20, 4:30~5:20, 5:30~6:20)

上課地點：明達 205 上課資料：講義 (請大家在每週上課前，來這個網頁把上課講義列印好)

上課內容將錄影並放置於 NTU Cool https://cool.ntu.edu.tw

丁老師聯絡方式 E-mail: jjding@ntu.edu.tw

繳交程式電子檔： NTUCool https://cool.ntu.edu.tw

Office：明達館723室, TEL： 33669652,

網頁：https://disp.ee.ntu.edu.tw/index.php

Office hour：我週一下午，週三上午，週四下午，週五下午都有空，歡迎同學們來找我

助教：曾子容，林政均，李沅罡，許宸睿，顏勁賢

，

上課資料專區

02月21日上課資料 ........PDF..........上課筆記檔(1)(pages 1~85)

03月06日上課資料 ........PDF

03月13日上課資料 ........PDF...........Homework 1 (3月20日繳交) ........PDF

03月20日上課資料 ........PDF..........上課筆記檔(2)(pages 86~180)

03月27日上課資料 ........PDF..........上課筆記檔(3)(pages 181~271)

04月03日上課資料 ........PDF...........Homework 2 (4月10日繳交) ........PDF

04月10日上課資料 ........PDF

04月17日上課資料 ........PDF..........上課筆記檔(4)(pages 272~333)

05月01日上課資料........PDF ...........Homework 3 (5月8日繳交) ........PDF

05月08日上課資料 ........PDF ..........上課筆記檔(5)(pages 334~399)

05月15日上課資料 ........PDF ...........Homework 4 (5月29日繳交) ........PDF

Note: Homework 4 第(4)題題目有修正

05月22日上課資料 ..........上課筆記檔(6)(pages 400~465)

06月05日上課資料 ........PDF ...........上課筆記檔(7)(pages 466~534

Homework 5 (6月19日繳交) ........PDF

　

口頭報告專區

(第一場) Gray Image Compression ...........報告投影片

(第二場) A Data Compression and Storage Optimization Framework for IoT Sensor Data in Cloud Storage...........報告投影片

(第三場) Large Audio Language Models ...........報告投影片

(第四場) Prompt Singer ...........報告投影片

(第五場) 生成式語音語言模型介紹 ...........報告投影片

(第六場) Enhancing Signal-to-Noise Ratio in THz Spectroscopy for Biomedical Applications ...........報告投影片

(第七場) Benchmarking Code-switching Abilities of Speech Foundation Models ...........報告投影片

(第八場) Personalization ...........報告投影片

(第九場) Indoor Millimeter-Wave Imaging Based on Sparsity Estimated Compressed Sensing and Calibrated Point Spread Function ...........報告投影片

(第十場) Analyzing the Impact of Adversarial Patterns on Neural Networks Using ADSP Techniques ...........報告投影片

　

　

公告：

(1) 目前口頭報告，tutorials，程式編寫名額皆已滿

若要編輯維基百科，請先向老師登記並且告知要編輯的兩個條目

(2) 關於作業程式題，請注意 (i) 要交 *.m 或 *.py 檔

(ii) 一般題目的答案，和程式碼應該要分成兩個檔案交，不宜用壓縮檔，以方便助教批改

(3) 上課影片可以由 NTU Cool 下載

(4) 上課問答題目就在影片中，學號尾數不同的要回答的問題也不同

上課問答的答案，請依學號寫在作業的最後面

　

　

Tutorials 專區

(Part 1: Signal Processing Theories)

Non-Linear Time Variant System Analysis

Structural Similarity

Quaternion Algebra

Nonuniform Sampling

低功耗SVD處理器設計

Compressive Sensing

Mellin Transform

Independent Component Analysis

Singular Value Decomposition (SVD)

Graph Signal Processing

Graph Fourier Transform

Golden-section Search and Fibonacci Search

Quantum Signal Processing

Fuzzy Logic in Signal Processing

Signal Processing for Big Data

Bargmann Transform

Gaussian Mixture Model

Hidden Markov Model

Sparse Representation

　

(Part 2: Filter Design)

Adaptive Filter Algorithm and Application

Bilateral Filter

Recent Development of Filter Design

Image Enhancement, Denoising, and Rain Removal

Particle Filter

Nonlinear Filter

Hamming and Hanning Filter

Hamming and Hanning Filter (1)

Butterworth and Chebyshev Filters

Learning Based Filters

Phase Noise Estimation Technique

　

(Part 3: Compression)

JPEG for Still Image Compression

Image Compression

Recent Development of Image Compression

Video Compression

Recent Development of Video Compression

H.264 Basic Architecture

Sound Signal Compression

Music Signal Compression

Text Compression

Arithmetic Coding

JBIG and Binary Image Compression

Medical Image Compression

Lossless Image Compression

Electrocardiogram Signal Compression

H.265 Video Compression Architecture

AV1 Video Coding

MP3 for Vocal Signal Compression

Sparse Coding and Dictionary Learning

Learning Based Image Compression Techniques

Learning Based Video Compression

High Efficiency Video Coding (HEVC)

Karhunen-Loeve Transform (KLT)

　

(Part 4: Music and Speech)

Automatic Music Genre Classification

Speech Analysis by Mel-Cepstrum

Music Emotion Recognition(1)

Music Emotion Recognition(2)

Query by Humming System(1)

Query by Humming System(2)

Mel-Frequency Cepstrum

Speaker Recognition

Speech Recognition in Multi-Speaker Scenario

Speech Emotion Recognition

Blind Source Separation

Natural Language Processing and Speech Processing

Distance Estimation for Sound Source

Echo Cancellation

Hearing Aid System

Audio Event Detection

Hearing-Aid Speech Quality Index (HASQI)

Transformer Models for Audio Signal Analysis

Transformer in Natural Language Processing

Automatic Music Evaluation

　

(Part 5: Optimization and Machine Learning)

Support Vector Machine

Vector Quantization

Optimization for L0 Norm Problems

Maximum Likelihood Estimation

Low Rank Approximation

Alternating Direction Method of Multipliers (ADMM)

　

(Part 6: Pattern Recognition)

Facial Expression Recognition

Face Recognition

Face Recognition(2016)

Face Detection and Adaboost

Occluded Face Recognition

Compressive Sensing for Pattern Recognition

Character Identification

Motion Recognition

Finger Tracking and Its Applications

Object Tracing for Videos

Vehicle Sensing and Tracking

　

(Part 7: Image Quality Improvement)

Image Deblurring

Image Denoising

BM3D Image Denoising Method

Shadow Removal(1)

Shadow Removal(2)

Image Sharpness

Image Enhancement

Rain Removal for Images

Reflection Removal in Image

Damaged Signal Recovery

Sharpness Measurement for Images

Image Quality Assessment

　

(Part 8: Image Processing)

Color Coordinate Transformation

Light Field Camera

半色調影像處理

Region of Interest Extraction

Structural Similarity

Scaling Invariant Feature Transform (SIFT)

Speeded-up Robust Feature (SURF) Points

Edge Detection (1)

Edge Detection (2)

Digital Watermarking

Video Summarization

Corner Detection

Image Segmentation Using Superpixels

Video Segmentation Using Supervoxels

Semantic Segmentation

Instance and Panoptic Segmentation

SAR Image Processing

Stereoscopic Image Processing

Hyperspectral Image Analysis

Satellite Image Processing

Image Stitching

Augmented Reality

Application of Electrooculogram in Virtual Reality (VR)

Super Resolution

Learning Based Superresolution

Image Registration

Image Inpainting

Transformer Models for Image Processing

　

(Part 9: Fast Algorithm)

Integer Cosine Transform

Fast Algorithm for DCT(1)

Fast Algorithm for DCT(2)

Winograd FFT Algorithm

　

(Part 10: Integer Transform and Number Theoretic Transform)

Integer Transform

Jacket Transform

Fermat and Mersenne Numbers

Chinese Remainder Theorem

Primitive Polynomial

　

(Part 11: Communication)

Multiple-Input and Multiple-Output System Analysis

Modern OFDM Techniques

Recent Development of CDMA

Multiple Signal Classification (MUSIC)

Target Localization Technique

Machine Learning in Wireless Communication

Massive MIMO Systems (1)

Massive MIMO Systems (2)

Universal Filtered Multi-Carrier Techniques

　

(Part 12: Medical Signal Processing and Medical Image Processing)

EMG(肌電圖) Signal Analysis

DSP for Electrocardiography Analysis

Signal Processing for Electroencephalography

DSP for DNA Sequence Analysis

Dynamic Programming for Sequence Alignment

Assistive Technology for Color Blindness

Magnet Resonance Imaging and Analysis(1)

Magnet Resonance Imaging and Analysis(2)

Signal Processing for Healthcare

Photoacoustic Imaging

Photoplethysmography (PPG)

Electroretinogram (視網膜電圖)

　

(Part 13: Other Applications of Digital Signal Processing)

Cryptography(1)

Cryptography(2)

Multimedia Security

Advanced Multimedia Security Techniques

3-D Accelerometer Signal Analysis

Doppler Ultrasound Signal Analysis

Data Prediction

Signal Processing for Economical Data Analysis

Learning Based Prediction Techniques

Financial Signal Processing

Recent Advance in Fintech

Remote Sensing Techniques

Millimeter Wave Radar Systems

　

　

評分方式：

平時分數: 15 scores

基本分12分，各位同學皆可拿到

另外再根據上課回答問題加分(依學號，寫在作業上)，答對一次加 0.8 分

Homework: 60 scores

(5 times, 每 3 週一次，第 3n+1 週出作業，第 3n+3 週交作業)

(請自己寫，和同學內容相同，將扣 70% 的分數，就算寫錯但好好寫也會給 40~95% 的分數，
遲交分數打 8 折，不交不給分。不知道如何寫，可用 E-mail 和我聯絡，或於上課時發問)

　

Term paper 25 scores

方式有五種

(1) 書面報告 (10頁以上(不含封面)，中英文皆可，11或12的字體，題目可選擇和課程有關的任何一個主題
格式和一般寫期刊論文或碩博士論文相同，包括 abstract, conclusion, 及 references，
並且要分 sections，必要時有subsections。 References 的寫法, 可參照一般 IEEE 的論文的寫法 )

(2) Tutorial (和書面報告格式相同，但18頁以上，題目由老師指定，以清楚的介紹一個主題的基本概念和應用為要求，
選擇這個項目的同學，學期成績加 3 分)

(3) 口頭報告 (限五個人，每個人 15~40分鐘，題目可選擇和課程有關的任何一個主題，選擇這個項目的同學，學期成績加 2分)

(4) 編輯 Wikipedia (中文或英文網頁皆可，至少 2 個條目，但不可同一條目翻成中文和英文。限和課程相關者，自由發揮，
個人的貢獻加起來要 80 行以上，越有條理、有系統的越好，有意願者請向老師登記)

(5) 編寫程式以協助建立程式資料庫(選擇二個和這門課相關的題目(不可以和歷年 Homework 重覆，不可太瑣碎)，來編寫相關的程式，
程式用 Matlab 或 Python 編寫皆可，有意願者請向老師登記)

　

　