-
-
15 November: The final report deadline has been extended to December 9th, 2019.
-
11 November: The deadline for Assignment 2 has been shifted to midnight on Monday, Nov. 18th.
9 November, 21:30: Assignment 2
PDF updated to fix typo in Equation 1.
7 November: Assignment 2 clarifications have been added to the
PDF concerning tie values in saliency maps, how to explain any included heuristics for ROI searches, and making clear the requirements for plots in Part 2 and Part 3 to be over all provided images, not produced from a single image.
5 November, 15:33: Assignment 2 updated - Part 2 threshold should vary from 10 to 90 in increments of 10, not 0.1 to 0.9 in increments of 0.1. Part 3 threshold was unspecified; this should also vary from 10 to 90 in increments of 10. The assignment
PDF has been updated to reflect this.
4 November, 14:57: Assignment 2 files posted for download at the bottom of the Schedule page.
4 November, 14:54: Please note that several typos were fixed on the
PDF for Assignment 2, most notably the inputs for the function in Part 1. This function does not depend on the original image, and only requires the saliency map. Please re-download the
PDF if you downloaded it earlier.
28 October: Labs this week (today and Wednesday, 30 October) will be dedicated time to work on projects to get ready for the site visits next week.
24 October: Demos posted from last lecture bundled with the forgotten Boundary Problem demo.
-
-
-
11 October, 10:22: Note that the assignment solutions have been updated to correct an error in the explanation for Question 2(b.) and for Question 3(c.).
-
7 October: The lab this week will be a catch-up lab for previous weeks' labs. The MNIST lab will take place following Reading Week.
2 October: I forgot to upload the saliency demo from Monday, so that is now bundled with the softmax demo from class today and posted to on the schedule page.
26 September, 11:00: A student noticed that the settings given for the output did not appear to indicate the correct thickness. I had an error in my script to generate the samples which lagged behind on the thickness variable; I apologize. I have updated the zip file with new outputs (note that I also changed the colours to indicate which image pairs are correct; if you are using up to date images your samples should be red with a thin blue border (sample1) and purple with a thick green border (sample2)).
26 September:
Question 4 samples - note that I misspoke in class today; the shadow font should only be the projection, but it is based on the input font as I described in class. EDIT: Note that the angle for this question is defined with 0 degrees lying in the positive x direction, and sweeping out in the clockwise direction.
25 September: New demo notebooks posted. Note that the steerable filters notebook is included with the corrected code for specifying kernel sizes.
23 September: Assignment 1 released. You can download it from the Schedule page.
18 September, 14:31: Demos from today's class have been posted under the lecture notes for today on the course Schedule page.
18 September, 14:04: Lecture slides from today updated to include the Elephant in the Room example (new final slide in the lecture) and to fix typos pointed out in class.
16 September, 14:40: Please note that the lecture slides from today (Image Representation 2) were updated to correct an error on slide 13, as well as to add the missing course announcements. Please note this change.
11 September: You can download the Jupyter notebook demos from today's lecture
here or on the
Schedule page. The Schedule has been updated with expected release dates for assignments and assignment and project deadlines.
Welcome to the Fall 2019 offering of Computer Vision! The first lecture has been posted under the
Schedule link in the left sidebar.