Type of Credit: Elective
Credit(s)
Number of Students
Taking a video and posting it on social media, sending video messages to friends, watching videos on Youtube, etc., have almost already become our daily lives. Without video compression, these would never become possible. This course offers students a fundamental understanding of video compression, covering the video coding flow and techniques, including the most commonly used video coding standards, H.264 and H.265. Also, we will also introduce some video compression or processing techniques based on deep learning.
能力項目說明
Students will be able to learn fundamental compression theories for images and videos. In addition, students will be required to implement some basic codec modules by themselves.
| 教學週次Course Week | 彈性補充教學週次Flexible Supplemental Instruction Week | 彈性補充教學類別Flexible Supplemental Instruction Type |
|---|---|---|
Week 1
Covering topics: Introduction
Reading: Chapter 1 in the textbook
Teaching/HW: Explaining the syllabus and introducing image processing
Hours spent for preview and review: 1 hour each
--
Week 2
Covering topics: Mathematical Background
Reading: Slides
Teaching/HW: Teaching fundamental math background for video compression
Hours spent for preview and review: 2 hours each
--
Week 4
Covering topics: Color space and video formats
Reading: Chapter 2
Teaching/HW: Introducing several color spaces and video formats often used
Hours spent for preview and review: 2 hours each
--
Week 5 & 6:
Covering topics: Predictive coding - spatial prediction
Reading: Chapters 3 and 7
Teaching/HW: Talking about how to use spatial information for prediction in order to compress data
Hours spent for preview and review: 2 hours each
--
Week 7: Watching online materials for AV1:
https://www.facebook.com/watch/live/?ref=watch_permalink&v=192587200575081
Week 8:
Covering topics: Predictive coding - temporal prediction (motion estimation and compensation)
Reading: Chapters 3 and 6
Teaching/HW: Speaking about how to use temporal information for prediction in order to compress data
Hours spent for preview and review: 2 hours each
Week 9: Midterm Exam
Week 10 & 11
Covering topics: Transform coding and quantization
Reading: Chapters 7 and 8
Teaching/HW: Teaching about how to transform data from the time domain to the frequency domain and to quantize signals for lossy compression
Hours spent for preview and review: 2 hours each
Week 12-13
Covering topics: In-loop filter and entropy coding
Reading: Chapters 8 and 9
Teaching/HW: Teaching about how to remove blocking artifacts in the encoding/decoding process and to turn the compressed data into bitstream (code)
Hours spent for preview and review: 2 hours each
Week 14
Covering topics: Rate-distortion optimization
Reading: Chapter 10
Teaching/HW: Explaining the trade-off between the coding bitrate and video quality and how to optimize them.
Hours spent for preview and review: 2 hours each
Week 15: Deep Learning Basics
Teaching/HW: Introduction to deep learning
Hours spent for preview and review: 2 hours each
Week 16: Flexible Teaching (H.264/H.265, Deep learning-based video compression processing)
Week 17: Preparation for the final project
Week 18: Final Presentation
Textbook:
Iain Richardson, “Video Codec Design: Developing Image and Video Compression Systems,” Wiley, 2002
Reference:
1. John Watkinson, “MPEG Handbook,” Focal Press, 2001
2. Gary J. Sullivan et al., “Overview of the High Efficiency Video Coding (HEVC) Standard,” IEEE TCSVT, 2012.
1. https://aomedia.org/ 2. https://hevc.hhi.fraunhofer.de/