Computer vision is a field that harnesses the power of computers to extract information of interest from visual data captured by sensors. This course offers an introduction to techniques for understanding the shapes, motions, and meanings of objects depicted in images. Its goal is to lay the groundwork for exploring advanced topics in computer vision, such as AI-driven image recognition and image generation.
- To be able to explain and implement methods of filtering, 3D reconstruction, and object tracking.
- To be able to explain methods of image retrieval, image recognition, and their applications.
✔ Applicable | How instructors' work experience benefits the course |
---|---|
The instructors have conducted R&D about computer-vision technologies such as 3D reconstruction, object tracking, image retrieval, image recognition, and image segmentation in the automotive industry. |
Filtering, Optical Flow, Epipolar Geometry, Stereo Matching, SLAM, Robust Estimation, Camera Calibration, Kalman Filter, Particle Filter, Principal Component Analysis, Singular Value Decomposition, Subspace Method
✔ Specialist skills | Intercultural skills | Communication skills | Critical thinking skills | Practical and/or problem-solving skills |
Instructor will use slides and sample programs in the lectures.
Course schedule | Required learning | |
---|---|---|
Class 1 | Image Processing | Image Acquisition, Geometric Transformation, Resampling, Encoding |
Class 2 | Filtering | Spatial Filtering, Spectral Filtering, Template Matching |
Class 3 | Programming: Filtering | Implementation of filtering with MATLAB |
Class 4 | 3D Reconstruction (1/3) | Optical Flow, Projective Transformation, Singular Value Decomposition, Epipolar Geometry |
Class 5 | 3D Reconstruction (2/3) | Factorization Method, Rectification, Robust Estimation |
Class 6 | 3D Reconstruction (3/3) | Bundle Adjustment, Camera Calibration |
Class 7 | Programming: 3D Reconstruction | Implementation of 3D reconstruction algorithms with MATLAB |
Class 8 | Object Tracking | Kalman Filter, Particle Filter |
Class 9 | Programming: Object Tracking | Implementation of object tracking algorithms with MATLAB |
Class 10 | Image Retrieval | Local Features, Approximate Nearest Neighbor Search |
Class 11 | Image Recognition | Subspace Method, Support Vector Machine, Feature Aggregation |
Class 12 | Vision for Autonomous Driving | Recognition, Path Planning, Multi-Task Learning, Reducing Computational Load |
Class 13 | Event-Based Vision | Odometry Estimation, SLAM |
Class 14 | Physics-Based Vision | Optical Properties of Object, Photometric Stereo, Shape from Shading |
Students are encouraged to review the materials taught in each class for about 1 hour.
None required.
Lecture slides are used during the class.
Richard Szeliski, Computer Vision: Algorithms and Applications, 2nd ed., Springer, 2011.
Richard Hartley and Andrew Zisserman, Multiple View Geometry in Computer Vision, 2nd ed., Cambridge University Press, 2004.
Presentation video (100%)
Students are required to have undergraduate-level knowledges on computer science, linear algebra, calculus, probability, and statistics.
isato[at]c.titech.ac.jp
17:05-17:20 on Tuesdays
It is recommended to set up an environment to use MATLAB.