Upload
others
View
14
Download
0
Embed Size (px)
Citation preview
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
CS576: Computer Vision
Min H. KimKAIST School of Computing
General course introduction
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Welcome
• [CS576] Computer Vision
• Professor: Min H. Kim (김민혁)([email protected])
• (Lecture) Monday and Wednesday 1:00PM—2:30PM, Rm. 117, Bldg. N-1
• Course website: http://vclab.kaist.ac.kr/cs576/
2
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
About Instructor• Min Hyuk Kim (김민혁)– Associate Professor at KAIST School of Computing since 2012
– Leading the Visual Computing Lab. (KAIST VCLAB) in School of Computing
– PhD in Computer Science (University College London, UCL 2010)
– Postdoc Researcher at Yale University
• Research Areas: Computer Graphics and Vision– Computational Imaging & Computational Photography
– Image-Based Modeling and Rendering
– Color & Visual Perception
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
About Instructor• Min Hyuk Kim (김민혁)– ACM SIGGRAPH Asia Technical Paper Committee (TPC) 2017, 2018– Eurographics Technical Paper Committee (TPC) 2018, 2019– Associate Editor of ACM Trans. Graphics (TOG) 2014--2017– Associate Editor of ACM Trans. Applied Perception (TAP) 2016--
Present– KAIST Top 10 Research Accomplishments 2016– Microsoft New Faculty Award 2013– Naver Young Faculty Fellowship 2015
4
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Textbook• Reference book: Richard
Szeliski (2010) Computer Vision: Algorithms and Applications, Springer.
• PDF version is available from SpringerLink within the KAIST network.
• https://link.springer.com/book/10.1007%2F978-1-84882-935-0
5
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Timeline
6
Low-level computer
vision(geometry, 3D
imaging)
High-level computer
vision (detection,
recognition)
Computational photography
(HDR and light-field imaging)
Machine-learning vision (CNN, GAN, RL,
etc.)
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Notification• There are no official course prerequisites.
• [Warning] However, we assume in CS576 – We do not teach general mathematics such as
linear algebra and calculus.
– Homework assignment will require programming experience in MATLAB and Python!
– We do not provide basic programming skills.
7
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Grading
• Class participation: 10%
• Midterm/final exam: 50% (25% each)
• Programming assignments (about 4 HWs): 30%
• Quizzes for every lecture: 10%
• An assignment after its original due date will be degraded from the marked credit per day:e.g., A à B for one-day delayed submission.
8
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
What we will do in this course• Checking your attendances for every lecture
• Quiz at the beginning of every lecture!
• Homework plagiarism– First, no grade for copied codes (such as your friends or
codes on Internet) is given! – Second, you will get F accordingly!
Your cheating will be reported to the school.– Do not cheat by copying others’ codes!– Our TA are using an advanced code plagiarism detection
software!• Midterm and final term examinations (all about math in
computer vision, not open-book, no memory test!)
• Let’s have fun! J9
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Teaching Assistants• Sukjun Jeon (VCLAB)– [email protected]
• Hyunho Ha (VCLAB)– [email protected]
• Hyeonjoong Jang (VCLAB)– [email protected]
• Shinyoung Yi (VCLAB)– [email protected]
• Inseung Hwang (VCLAB)– [email protected]
10
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
INTRODUCTION TO COMPUTER VISION
11
Acknowledgement: many slides here are borrowed from Professor James Tompkin at Brown University, USA, and Reinhard Klette at Auckland University of Technology, New Zealand.
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
What is computer vision?Have you ever used computer vision?How? Where?
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
What is computer vision?Jitendra Malik, UC BerkeleyThree ‘R’s of Computer Vision
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
What is computer vision?Jitendra Malik, UC BerkeleyThree ‘R’s of Computer Vision
“The classic problems of computational vision:reconstruction (in 3D)recognition(re)organization.”
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
What is computer vision?Have you ever used computer vision?How? Where?Reconstruction? Recognition? (Re)organization?
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Optical character recognition (OCR)
Digit recognition, AT&T labs
http://www.research.att.com/~yann/
Technology to convert scanned docs to text
• If you have a scanner, it probably came with OCR software
License plate readers
http://en.wikipedia.org/wiki/Automatic_number_plate_recognition
JH
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Face detection
• Almost all digital cameras detect faces• Snapchat face filters
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Smile detection
Sony Cyber-shot® T70 Digital Still Camera JH
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Object recognition (in supermarkets)
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
How does it work?
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Vision-based biometrics
“How the Afghan Girl was Identified by Her Iris Patterns” Read the story (Wikipedia)
JH
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Login without a password…
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Login without a password…
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Login without a password…
Liang et al. 2014
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Object recognition (in mobile phones)
e.g., Google Lens
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
3D from images
Building Rome in a Day: Agarwal et al. 2009
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Human shape capture
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Human shape capture
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Human shape capture
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Human shape capture
Lecturer: Min H. Kim (KAIST) CS576: Computer VisionStar Wars: Rogue One – Peter Cushing / Admiral Tarkin
Special effects: shape capture
http://cinetropolis.net/tarkin-care-of-business-rogue-ones-digital-peter-cushing/
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Special effects: shape capture
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Special effects: motion capture
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Interactive Games: Kinect• Object Recognition:
http://www.youtube.com/watch?feature=iv&v=fQ59dXOo63o• Mario: http://www.youtube.com/watch?v=8CTJL5lUjHg• 3D: http://www.youtube.com/watch?v=7QrnwoO1-8A• Robot: http://www.youtube.com/watch?v=w8BmgtMKFbY
JH
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Sports
Sportvision first down lineNice explanation on www.howstuffworks.com
http://www.sportvision.com/video.html
JH
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Medical imaging
Image guided surgeryGrimson et al., MIT3D imaging
MRI, CT
JH
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
AutoCars - Uber bought CMU’s lab
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Industrial robots
Vision-guided robots position nut runners on wheels
JH
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Vision in space
Vision systems (JPL) used for several tasks• Panorama stitching• 3D terrain modeling• Obstacle detection, position tracking• For more, read “Computer Vision on Mars” by Matthies et al.
NASA'S Mars Exploration Rover Spirit captured this westward view from atop a low plateau where Spirit spent the closing months of 2007.
JH
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Mobile robots
http://www.robocup.org/NASA’s Mars Spirit Rover
http://en.wikipedia.org/wiki/Spirit_rover
Saxena et al. 2008STAIR at Stanford
JH
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Augmented Reality and Virtual Reality
MS HoloLens, Oculus, Magic Leap,ARCore / ARKit
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
What is computer vision?
Jitendra Malik, UC BerkeleyThree ‘R’s of Computer Vision
“[Further progress in] the classic problems of computational vision:
reconstructionrecognition(re)organization
[requires us to study the interaction among these processes].”
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
History• Origins in Early 1960s
• First Textbook– Azriel Rosenfeld. Picture Processing by Computer.
Academic Press, New York, 1969.
• First Journal– Computer Graphics and Image Processing (1972
– 1982) founded by Azriel Rosenfeld, – after 1982 continued as Computer Vision, Graphics,
and Image Processing48
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Conference history• Int. Conf. Pattern Recognition (ICPR)– Washington D.C. 1973, chair: King-Sun Fu
• IEEE Conf. Computer Vision Pattern Recognition (CVPR)– Arlington 1983, chairs: Takeo Kanade and Dana Ballard
• Int. Conf. Computer Vision (ICCV)– London 1987, chairs: J. M. Brady and Azriel Rosenfeld
• European Conf. Computer Vision (ECCV)– Antibes 1990, chair: Olivier Faugeras
• Asian Conf. Computer Vision (ACCV)– Osaka 1993, chair: K. Sugihara
49
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
State of the art today?With enough training data, computer vision nearly matches human vision at most recognition tasks
Deep learning has been an enormous disruption to the field. Many techniques are being “deepified”.
JH
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Computer Vision and Nearby FieldsDerogatory summary of computer vision:
“Machine learning applied to visual data.”
JH
Lecturer: Min H. Kim (KAIST) CS576: Computer Vision
Computer Vision and Nearby Fields
Derogatory summary of computer vision:“Machine learning applied to visual data.”
JH
Model of the world
Images, videos,sensor data…
Images, videos,interaction
Digital worldReal world
Computer Graphics Computer Vision
Question answering