Instructors
Overview
Introduction to the principles and practice of computer vision. This course will provide in-depth survey of topics involved in major tasks in modern vision, and offer hands-on experience in implementing some of them.
Prerequisite
A good background in linear algebra and probability (undergrad level) and background in machine learning (TTIC 31020 or equivalent) required. CMSC 25040 is recommended.
Topics
- Image formation, representation, and compression
- Representation and perception of color
- Filtering and edge detection
- Image features, detectors, and interest point operators
- Model fitting, RANSAC and Hough transform
- Stereo and multi-view geometry
- Camera calibration
- Representation and perception of motion
- Representation and modeling of edges and regions
- Semantic vision: recognition, detection, and related tasks
Expected Outcomes
- Familiarity with models for of image formation and image analysis.
- Familiarity with major methods for image processing, alignment, and matching.
- Knowledge of principles and methods of 3D reconstruction from images and videos, and ability to build and diagnose implementations of these methods.
- Grasp of modern methods for object and scene categorization from images, and ability to build and diagnose implementations of such methods.
Time and Venue
Lectures on Mon, Wed 1:30pm - 2:50pm, at TTIC 523. First two weeks will be remote. Afterwards we will resume in-person instruction.
Schedule
| WEEK | DATE | EVENT | MATERIALS |
|---|---|---|---|
| week 1 | Mon 01-10 | General intro; what is [computer] vision? | |
| Mon 01-10 | Assignment 1: Warmup | ||
| Wed 01-12 | Camera models | ||
| week 2 | Mon 01-17 | MLK day break | |
| Wed 01-19 | Single view metrology; Homographies | ||
| Fri 01-21 | Assignment 2: Rendering, Homography and Calibration | ||
| week 3 | Mon 01-24 | Camera Calibration | |
| Wed 01-26 | SfM Intro: Triangulation and Epipolar Geometry | ||
| week 4 | Mon 01-31 | Structure from Motion | |
| Wed 02-02 | SfM; Stereo Matching | ||
| Sun 02-06 | Assignment 3: Structure from Motion | ||
| week 5 | Mon 02-07 | Depth from Stereo and Mono | |
| Wed 02-09 | Photometry | ||
| week 6 | Mon 02-14 | Photometric Stereo; Intro to Filtering | |
| Wed 02-16 | Image Filtering (bilateral, non-local, etc) | ||
| week 7 | Mon 02-21 | Edge Detection | |
| Wed 02-23 | Hough Transform, RANSAC, Keypoints | ||
| Fri 02-25 | Assignment 4: Photometry and Feature Matching | ||
| week 8 | Mon 02-28 | Scale-space Keypoint Detectors; SIFT, Harris-Laplace | |
| Wed 03-02 | Recognition; Object Detection | ||
| Fri 03-04 | Project | ||
| week 9 | Mon 03-07 | Motion and Optical Flow | |
| Wed 03-09 | Color; Multiview Stereo and NeRF | ||
| week 10 | Mon 03-14 | Reading Period | |
| Wed 03-16 | End of Quarter |
Made with ❤ by Yours Truly.