智慧光電影像處理

智慧光電影像處理(Intelligent Photonic Image Processing)以於空間域光電影像處理為主軸,搭配深度學習於影像辨識,先藉由連續與離散傅立葉轉換(Fourier transform)來做光電影像分析以及在光電訊號系統上應用,並著重理論與實作的搭配,使修課的學生能真正的了解連續與離散的光電影像處理之原理與系統。利用這些原理基礎來實作單元一(二維連續空間):光學特徵辨識系統(Lab 1);單元二(二維離散空間):適應性濾波器之影像強化技術(Lab 2) ;單元三(影像智慧辨識):藉由Convolutional Neural Networks之影像辨識技術(Lab 3)。期使學生能夠熟知光電影像處理之原理與系統,以實驗應證這些光電影像的處理;進而運用這些光電影像處理理論與技術,來整合與發展智慧光電影像系統工程。 課程內容: 1. Optical Information Processing: (2D Continuous-Space Signal Processing) 6 weeks a. 2D Signals and Systems: Analysis, Scalar diffraction theory, and Fresnel and Fraunhofer diffractions. b. Optical Information Systems and Processing: Wave-optics analysis of coherent optical systems, Frequency analysis of optical imaging systems, and Analog optical information processing. Lab 1: Optical Character Recognition (OCR) System: Matched Filter and Optical Decorrelation Ref.: J. W. Goodman, Introduction to Fourier Optics, 3rd Ed., Robert & Company, 2005. R. Bracewell, The Fourier Transform & Its Applications, 3rd Ed., McGraw-Hill, 1999. 2. Digital Image Processing: (2D Discrete-Space Signal Processing) 6 weeks a. 2D Signals and Linear Systems: Mathematical preliminaries (point spread function), Image perception, Image sampling & quantization. b. Image Transforms: Matrix operations, Random signals, Unitary transforms, Discrete Fourier transform (DFT), Discrete cosine transform (DCT), Karhunen-Loeve (KL) transform, Singular value decomposition (SVD), and Wavelet transform. c. Image Enhancement and Restoration: Histograms, Edge detections, Inverse filters, Wiener filter, and Adaptive Wiener filter. Lab 2: Image Filtering and Restoration with Adaptive Wiener Filter Ref.: A. K. Jain, Fundamentals of Digital Image Processing, Prentice-Hall, 1989. R. C. Gonzalez and R. E. Woods, Digital Image Processing, 4th Ed., Pearson Education, 2018. MATLAB with Image Processing Toolbox. 3. Image Recognition with Deep Learning: 6 weeks a. Image Segmentation: Edge Detection, Thresholding, Region Detection, Active Contours, and Feature Extraction. b. Image Pattern Classification: Patterns and Pattern Classes, Neural Networks and Deep Learning, and Deep Convolutional Neural Networks. Lab 3: Image Recognition with Convolutional Neural Networks (Matlab) Ref.: R. C. Gonzalez and R. E. Woods, Digital Image Processing, 4th Ed., Pearson Education, 2018. MATLAB with Deep Learning Toolbox.

任課教師:陳顯禎 @ 5F R520;電話:(06) 3032121 ext. 57825或0931178072;E-mail: sheanjen@nctu.edu.tw 課程助教:王基祐;電話:(06) 3032121 ext. 57807;E-mail: chiyu9015@gmail.com 交大光電學院適應性光子實驗室(Adaptive Photonics Lab.) @光電學院4F R440 備註: 上課時間:週四 14:20~17:10 (無實驗週) 週四 14:20~16:10 (有實驗週) 實驗時間:週四 18:30~21:20 上課地點:奇美樓CM217教室 實驗地點:奇美樓R219光電系統教學實驗室 & R224電腦模擬教室

成績計算:總分110分,超過100分以一百分計算 1. 作業(每單元2次,共6次):4×6=24 2. 實驗報告(每單元1次,共3次):12×3=36 3. 期中考試:25 4. 期末考試:25

J. W. Goodman, Introduction to Fourier Optics, 3rd Ed., Robert & Company, 2005. R. Bracewell, The Fourier Transform & Its Applications, 3rd Ed., McGraw-Hill, 1999. A. K. Jain, Fundamentals of Digital Image Processing, Prentice-Hall, 1989. R. C. Gonzalez and R. E. Woods, Digital Image Processing, 4th Ed., Pearson Education, 2018. MATLAB with Image Processing Toolbox. R. C. Gonzalez and R. E. Woods, Digital Image Processing, 4th Ed., Pearson Education, 2018. MATLAB with Deep Learning Toolbox.