Due: Late (*) Sunday, October 21.
(*) "Late" means that the
instructor might start grading by 8AM Monday. Once the instructor starts
grading, no more assignments will be accepted.
5 points
This assignment should be done individually
You can do the programming parts of this assignment in any language you like,
although if your results are anomalous, and the grader does not speak the
language you use, they may be less able to quickly figure out what the problem
is and give you reasonable part marks.
Regardless of what language you use, please follow the instructions linked here carefully.
Deliverables
within questions are flagged with ($).
This
assignment has two questions. The first regular problem is worth 3 points, and
the second one is worth 2, for a total of 5.
The image pair
are an image of a PowerPoint slide and
a frame from a video of the lecture using that slide.
i) (a) Select 4 pairs of corresponding
points in the two images. For example, you may want to use the corners of the
slide. Call this set (A,A’) where A is the (x,y) for 4 points on the slide
image, and A’ is the (x’,y’) for 4 points on the
frame image. Report your points and how you got them ($). Repeat this for 4
more pairs of points, (B,B’), and again report your points and you got them
($),
i) (b) Add squares of differing colors
(for each pairs) to the images for both (A,A’) and (B,B’). (You can build on
code that you should already have from a previous assignment). Put the two
images into your report with an informative caption ($).
ii) (a) The Matlab function direct_linear_transform
(linked) takes two matrices with 2 columns and the same number of rows
(there must be at least four) computes the homography mapping, H, that maps the
first one to the second. Use that function to compute the homography that maps
the slide points in A to frame points in A’. Report the homography, H ($). Now
use H to compute both where the slide points in A map to in the frame image,
and similarly where B are mapped into the frame image. Visualize this by adding
squares onto the modified frame image that shows your selected points (they are
the “target”). Put this image into your report with an informative caption ($).
ii) (b) We can measure the
effectiveness of the mapping by computing the average Euclidean distance
between the mapped points (i.e., where H tells the points in A to go) and the
selected points (i.e., the points in A’). Compute this distance for both sets
of four points. Report the two error measures ($). Comment on whether the
mapped points from A are closer to the target points than the mapped points
from B or vice versa (you can also refer to your image to discuss this rather
than just relying on what a single number tells you). If there is a difference
(generally there will be one), offer an explanation ($).
iii) (a) By some process (perhaps by
selecting 2 points) define a rectangle to put a box around the title of the
slide, and a second one to put a box around the bullet points. Use the line
drawing capability from a previous assignment to draw these to boxes in red on
the slide image. Put the resulting image into your report with an informative
caption.
iii) (b) Using the homography
transformation already computed, “improve” the slide frame by putting a
corresponding box (this time green) into the frame image.
Consult
the instructions linked here for conventions for preparing and handing
in assignments. In 2012, hand in assignments via email to Kyle Simek
(ksimek@email.arizona.edu).