Digital Signal Processing
| Numbering Code | U-ENG26 36061 LJ72 | Year/Term | 2022 ・ Second semester |
|---|---|---|---|
| Number of Credits | 2 | Course Type | Lecture |
| Target Year | Target Student | ||
| Language | Japanese | Day/Period | Mon.4 |
| Instructor name |
NOBUHARA SHOUHEI (Graduate School of Informatics Associate Professor) NISHINO KO (Graduate School of Informatics Professor) |
||
| Outline and Purpose of the Course | The goal of this lecture is to understand fundamental theories and filter designs for one-dimensional time-domain signal and two-dimensional image processing and encoding. In particular, this course provides introductions to orthogonal transformation such as discrete Fourier transform, fast Fourier transform algorithms, one-dimensional and two-dimensional signal encoding methods including basics of JPEG / MPEG, and FIR and IIR filters based on the discrete-time linear time invariant system theory | ||
| Course Goals | Digital signal processing requires both theoretical analysis / design and practical software system implementations. This course provides exercises on signal processing in Python, with guidance by teaching assistants, and additional resources via the course web site. Short questions and answers are also provided to help understand the theories and implementations. | ||
| Schedule and Contents |
Overview of digital signal processing (2 classes) * Introduction of the goal of digital signal processing, its essential ideas and advantages. * Extension of 1D Fourier transform to 2D or multi-dimensional signals and its applications in computed tomography (CT). Sampling and quantization (1 class) * Sampling theories in 1D signals and digitization process of 2D images. Discrete Fourier transform and FFT (3 classes) * Discrete Fourier transform in 1D signals. * Fast Fourier transform and its extension to 2D image signals. Orthogonal transformation and short-time Fourier transform (3 classes) * Discrete cosine transform and digital signal processing based on orthogonal transformation. * Short-time Fourier transform. * Multi-scale signal analysis and its extension to wavelet transform. Encoding (2 classes) * Waveform coding, vector quantization, and transform coding. * Media encoding for audio, document images, images (JPEG) and videos (MPEG). Filtering based on discrete-time systems (3 classes) * Discrete-time linear time-invariant system and z transform. * FIR and IIR filters * Basics on linear phase FIR filter and IIR filter design. * Filtering of 2D image signals. |
||
| Evaluation Methods and Policy | Grade evaluations will be based fundamentally on scores in the written final test. Evaluation will also be provided for the development of “non-trivial” digital processing software and reports on the functions, designs, performance evaluations, etc., of the software. | ||
| Course Requirements |
Industrial Mathematics E1 (20540) and Fundamental Communication Theory (60320) are prerequisites for this course. Students should take Digital Control (60270) in parallel. |
||
| Study outside of Class (preparation and review) | Students should improve programming skills in Python through digital signal processing exercises provided in the lecture. | ||
| References, etc. | Think DSP: Digital Signal Processing in Python, Allen B. Downey, (O'Reilly Media), ISBN:1491938455, Online versions are available at http://greenteapress.com/wp/think-dsp/ | ||
| Related URL | http://greenteapress.com/wp/think-dsp/ | ||
