Computational Imaging explores the joint design of imaging hardware and computational algorithms to capture and reconstruct visual information. The course connects digital photography, optics, human perception, and inverse problems with modern learning methods and emerging imaging systems. Students develop practical understanding through four assignments that combine hardware and software.

By the end of the course, students will be able to:
1. Explain image formation and the roles of lenses, sensors, image signal processing, and human perception.
2. Formulate imaging tasks as forward and inverse problems and understand optimization-based and learned reconstruction methods.
3. Implement and evaluate computational imaging methods using Python, and interpret their assumptions and limitations.
4. Describe the principles of imaging and display techniques covered in the course, including multi-view, spectral, event, time-of-flight, polarization, and wave-optical methods.
5. Explain and defend the methods, implementation, and results of their work through demonstrations, presentations, and questions and answers.

Students are expected to have a working knowledge of Python programming and linear algebra. Experience with computer graphics, computer vision, machine learning, optics, or signal processing is helpful, but is not required.

Optional references and companion learning resources:

1. Bhandari et al., Computational Imaging, MIT Press, 2022.
https://imagingtext.github.io/
2. Stanford EE367: Computational Imaging.
https://web.stanford.edu/class/ee367/
3. Carnegie Mellon 15-463: Computational Photography.
https://imaging.cs.cmu.edu/15-463/
4. University of Toronto CSC2529: Computational Imaging.
https://www.cs.toronto.edu/~lindell/teaching/2529/
5. Columbia University: First Principles of Computer Vision.
https://fpcv.cs.columbia.edu/
6. 5 Minutes Graphics.
https://g5m.cs.washington.edu/

Fall 2026. The schedule is tentative and subject to change. All assignment deadlines are 23:59 KST.

Week 1
Sep 8: Introduction to computational imaging.
Sep 10: Digital photography: lens. Assignment 1 released.

Week 2
Sep 15: Digital photography: sensor.
Sep 17: Digital photography: ISP.

Week 3
Sep 22: Human perception.
Sep 24: Chuseok Holiday — no class.

Week 4
Sep 29: Calibration.
Oct 1: High dynamic range imaging. Assignment 1 due; Assignment 2 released.

Week 5
Oct 6: Inverse problem.
Oct 8: Inverse solver.

Week 6
Oct 13: Introduction to neural networks.
Oct 15: Learned inverse solvers — guest lecture by Prof. Hyungjin Chung (in person).

Week 7
Oct 20: Rendering.
Oct 22: Multi-view stereo and photometric stereo.

Week 8
Oct 27: Midterm Exam Week — no class. Assignment 2 due.
Oct 29: Assignments 1–2 demo and presentation.

Week 9
Nov 3: Appearance modeling and capture. Assignment 3 released.
Nov 5: Spectral imaging.

Week 10
Nov 10: Event sensing.
Nov 12: Time-of-flight imaging.

Week 11
Nov 17: Polarization imaging.
Nov 19: Wave optics for imaging.

Week 12
Nov 24: Unconventional imaging — guest lecture (speaker TBA). Assignment 3 due; Assignment 4 released.
Nov 26: Wave optics for imaging.

Week 13
Dec 1: Nanophotonics — guest lecture (speaker TBA).
Dec 3: Holographic display — guest lecture by Prof. Suyeon Choi (online).

Week 14
Dec 8: Alternative Holiday for the Foundation Day — no class.
Dec 10: Optical neural network.

Week 15
Dec 15: Structured light.
Dec 17: Light transport.

Week 16
Dec 22: Final Exam Week — no class. Assignment 4 due.
Dec 24: Assignment demo and presentation.

Class meetings: Tuesdays and Thursdays, 09:30–10:45, Engineering Building #2, Room 106.

The course consists of lectures, guest lectures, four hardware-and-software assignments, and assignment demonstrations and presentations.

Course website:
https://sites.google.com/view/shbaek/teaching/computational-imaging
The website provides course information and the schedule. Lecture slide and recording links will be added as they become available.

PLMS:
https://plms.postech.ac.kr/course/view.php?id=16248
PLMS is used for Q&A, assignment submissions, and course announcements.

Teaching assistant: Seungwoo Yoon (ysw1110@postech.ac.kr). Office hours are available by email appointment with the instructor or teaching assistant.

Attendance requirement:
Students must attend at least three quarters of all class hours to receive a grade.

Excused absences:
An excused absence counts as attendance only when approved in advance, or within seven days after the absence at the latest. Apply through the e-attendance system with supporting documents, contact the instructor, and confirm the approval result. Excused absences may not exceed one quarter of total class days.

Guest lecture attendance:
Attendance at all guest lectures is mandatory unless the student has an approved excused absence. Each unexcused guest lecture absence results in a 5% deduction from the overall final course score.

POSTECH rules and request form (Article 8; POSTECH login required):
https://board.postech.ac.kr/home/mainHome.do?boardNo=79

Study approach:
Review lecture materials and connect physical imaging principles with computational models. Use Python experiments to examine how modeling assumptions and reconstruction methods affect results. Consult the companion resources for additional explanations and examples, and use PLMS or class time to ask questions.

Collaboration:
Discussion with classmates is encouraged. Each student must submit their own solution and be able to explain it.

AI assistance policy:
AI assistants (e.g., ChatGPT, Claude, Copilot) are allowed for assignments. Students are responsible for understanding the work they submit. During the demo/presentation and Q&A, students must demonstrate their own understanding of the methods, implementation, and results.