EE · Spring 2026 · Undergraduate
Image Processing
From image formation and filtering to multiview geometry, visual recognition, and learning-based vision.
Course Overview
This course introduces the principles and algorithms that enable computers to process and understand visual information. It begins with image representation, filtering, and feature extraction, then develops the geometry of cameras and multiple views. The final part connects these foundations to image classification, neural networks, and convolutional models.
- Instructor
- Prof. Nuri Kim
- Term
- Spring 2026
- Meetings
- Mon 09:00–10:30
Wed 09:00–09:50 - Classroom
- Engineering Building 7
Room 114
Learning Objectives
Image processing and geometry
- Represent, filter, and analyze digital images in spatial and frequency domains.
- Detect corners and local features for matching visual observations.
- Model cameras, transformations, homographies, and two-view geometry.
- Recover depth and 3D structure from multiple images.
Recognition and learning
- Build image-classification pipelines using local and learned features.
- Explain neural networks and convolutional architectures for vision.
- Connect classical image representations to neural-network-based recognition.
- Implement and evaluate computer-vision methods in Python.
Instruction is in Korean, with an English-language textbook and technical terminology. Familiarity with Python and introductory linear algebra is recommended.
Schedule
Images & Features
Visual Geometry
Recognition & Learning
The schedule follows the Spring 2026 course plan. Slides, code, and detailed deadlines are available to enrolled students through the JBNU LMS.
Coursework
Assignment 0
Data Interpretation & Visualization
Explore numerical image data and build reliable visualizations.
Assignment 1
Filtering & Edge Detection
Implement convolution, image filtering, and edge extraction.
Assignment 2
Panorama Stitching
Detect and match features, estimate homographies, and warp images.
Assignment 3
3D Reconstruction
Estimate epipolar geometry and reconstruct structure from multiple views.
Assignment 4
Scene Recognition
Build a visual-words representation and evaluate scene classification.
Assignment 5
Neural Networks for Recognition
Implement and train neural models for visual recognition tasks.
- Assignments
- 30%
- Midterm Exam
- 20%
- Final Exam
- 40%
- Attendance
- 10%
Generative AI may be used creatively in course activities and assignments. Students remain responsible for checking, revising, and understanding all submitted work; unreviewed AI-generated submissions may be treated as academic misconduct.
Resources
- Course platformJBNU LMS
- Primary textbookRichard Szeliski, Computer Vision: Algorithms and Applications, 2nd ed.
- ReferenceRichard Hartley and Andrew Zisserman, Multiple View Geometry in Computer Vision
- ReferenceDavid Forsyth and Jean Ponce, Computer Vision: A Modern Approach
- ReferenceRafael Gonzalez and Richard Woods, Digital Image Processing
- Contactnuri.kim@jbnu.ac.kr