CS 453 - Spring 2016
Self-scheduled (pick up from my office door), Wednesday 5/18 - Tuesday 5/24.
The exam is 3 hours (but you have an extra 30 minutes if you need them),
open book / notes. You may use a computer
only to access our online textbook and our course page,
including homework, solutions, and student presentations.
You may not
use the computer for any other purpose, including searching the
You must return the exam within 4 hours after picking it up.
Record sign-out and sign-in time and date on the sheet on my office
door. Slide the completed exam under my office door.
All exams must be back by 5pm on Tuesday May 24.
Seniors must return their exams by 5pm on Sunday, May 22.
Practice questions on epipolar geometry.
- Homework 1, due Wednesday 2/24.
- Homework 2, due Wednesday 3/2.
- Homework 3, due Wednesday 3/9 and Friday 3/11.
- Homework 4, due Friday 3/18.
- Homework 5, due Friday 3/25.
- Homework 6, due Wednesday 4/6.
- Homework 7, due Friday 4/15.
Solutions + timing results.
- Homework 8, due Monday 4/25.
Stereo ranking results.
- Homework 9, due Monday 5/2.
- Final project
and list of projects.
Final reports due Monday, 5/16.
Lectures & Readings
- M 2/15 - Course info, introduction to computer vision (Sz 1; optical illusions)
- W 2/17 - Image formation, perspective projection, image library (Sz 2; HW 1;
- F 2/19 - Ideal vs. real images, image formats, run-encoding (SS 2.3-5;
- M 2/22 - Binary image analysis, filtering, connected components (SS 3.1-4; Sz 3.3.4)
- W 2/24 - Binary morphology (SS 3.5; Sz 3.3.2)
- M 2/29 - Noise, box filter, Gaussian smoothing (TV 2.3.3; Sz 3.2.3)
- W 3/2 - Edge detection, image gradients, difference operators (Sz 4.2.1; TV 4)
- F 3/4 - Animal vision (Felix),
HDR imaging (Joey), Laplacian and Marr-Hildreth edge detector (Sz 4.2.1; TV 4.2)
- M 3/7 - Canny edge detector, edge detection summary, Hough transform (Sz 4.2.1-2, 4.3.2; TV 4)
- W 3/9 - Hough transform, least-squares fitting (Sz 4.3.2, A2; TV 5.2; SS 10.3.4,
- F 3/11 - Image
matting (Peter), HW 4 code (Sz 10.2)
- M 3/14 - Color (Sz 2.3.2; SS 6.1-4)
- W 3/16 - Image stabilization (Jordan),
Feature detection and matching (Ben),
geometric transforms (Sz. 4.1.1-3, 2.1.2)
- F 3/18 - JPEG encoding (Kevin),
image warping, interpolation (Sz 5.1.1, 3.6.1)
- M 3/21 - Image compositing, homogeneous coordinates, homographies (Sz 3.1.3, 2.1.1-3)
- W 3/23 - 3D vision, projection matrix, intrinsic vs. extrinsic camera parameters (Sz 10.5, 2.1.5)
- F 3/25 - Camera calibration (Sz 2.1.5; TV 6.1, 6.3.1),
Line labeling (Shannon),
Texture synthesis (Dylan)
- M 4/4 - Exam info, camera calibration, SVD (TV 6.3.1, Sz A.1.1)
- W 4/6 - Exam review, intro to stereo (Sch 1.2, Sz 11.1)
- F 4/8 - Epipolar geometry, local correspondence, SSD (Sch 1.2, Sz 11.1, 11.3-4,
- M 4/11 - Stereo: SSD, SAD, NCC, disparity space (Sch 5.2, 6.1)
- W 4/13 - Middlebury Stereo Page,
robust matching, shiftable windows, adaptive weights, cross-checking,
- M 4/18 - Global matching, SGM stereo
(Sz 3.7.2, 5.5, 11.5, Middlebury MRF page)
- W 4/20 - Video - Human-centric image understanding
- F 4/22 -
Shape from shading (Zale);
stereo cross checking; structured light (Sz 12.1-2, stereo talk)
- M 4/25 -
Shape from defocus (Sebastian),
Feature tracking (Vaasu),
motion, optical flow (Sz 4.1.4, 5, 8.4; SS 9.3.4;
Middlebury flow page)
- W 4/27 - Tomasi and Kanade's SFM factorization method (Sz 7.3, A.1.1,
- F 4/29 - Intro to OpenCV
- M 5/2 - Final project info, optical flow, intro to recognition (Sz 8.1.3, 8.4, 14;
- W 5/4 -
Object recognition with CNNs (Max), more recognition
(Sz 14; UW CS576 slides:
- F 5/6 - Face detection (Sam), principal component matching, "Eigenfaces"
UW CS576 slides:
- M 5/9 - Snakes (John), epipolar geometry (Sch 1.2)
- W 5/11 - Seam carving (Dana),
Computational photography (slides by Rick Szeliski)
- F 5/13 - Final project demos
- M 5/16 - Course summary and final exam info
Books on 2-hour reserve:
Shapiro & Stockman (SS);
Trucco & Verri (TV);
Forsyth & Ponce;
Sonka, Hlavac, & Boyle.