« Lecture #Syllabus for all sect | Main | VisualStuSyllabus for all sect »
January 26, 2006
Project #Syllabus for all sect
The project file distribution (mirror at faculty.cs.byu.edu) is available.
1/27 at 9:24pm all future updates will appear as seperate entries in the projects category which will appear in the Projects Category Archive.
1/26 at 8:27pm I put up another file distribution, linked above, that has all the methods and classes that you'd expect from reading my comments in other places on the blog. This is not a critical update, just a useful one. If you don't want to use it, you don't have to. Note that this distribution actually includes some of the things you are supposed to implement yourself (like the segmentation point reader and gradient function). The due date isn't going to change (but remember that you have 5 late days per semester and upto 3 per project if you want to use them).
As of 1/23 at 2:50pm: We have posted a new (a quite probably final) file distribution. The file description has been updated to include more details on the simple greedy algorithm and an additional thing you have to turn in (it is just a copy of the code for your Dijkstra's implementation). In the project files, the main difference is that the gray images have correct segmentation points in the sense that the segmentation points now segment a meaningful part of the image. You can get just the pictures with the correct segmentation points if you like, or, you can get just the new project description without changes highlighted).
The project distribution includes a VS 2005 project and a project description. Basically, you are going to code up a couple ways to segment images given segmentation points and compare the algorithms in terms of correctness and running time. You will turn in a brief report. You need to do all this within a time bound set forth in the project description.
This is a draft in the sense that it contains all of the information we think is neccesary to complete the lab, but, may be changed later once the TAs have time to go through it and try it out.
Also, my pgm file reader is woefully slow and I am hoping that the TAs can Syllabus for al
Posted by jones at January 26, 2006 09:12 AM
Comments
I downloaded the project and noticed some things that could be improved on for the base project. I've got an update that I'm hoping to share for the benefit of the other students.
Dr. Jones, how can I show you the new project so you can review it?
Posted by: Andrew Arnott ![[TypeKey Profile Page]](http://vv.cs.byu.edu/cs312/nav-commenters.gif)
at January 18, 2006 02:36 PM
Andrew and Ryan the TA are working on weaving their changes. The interfaces to the objects will remain the same, but the implementations will be improved when I post their final project.
In general, I am entusiastic about student contributions to the class projects as long as they help make the class better, so keep them coming.
Posted by: Mike Jones at January 18, 2006 05:16 PM
If anyone is having trouble with Dijkstra's Algorithm or just wants to check if they are doing it correctly on an example. I found this cool website that alows you to input a sample graph and then steps you through Dijkstra's on the graph. Check it out!!
http://carbon.cudenver.edu/~hgreenbe/sessions/dijkstra/DijkstraApplet.html
Posted by: Cindy at January 26, 2006 01:31 AM
The skull-ascii-upper-jaw.pgm file has one errant segmentation point that makes the Dijkstra reach for a point way off course. Here is the corrected series of points:
(214, 348); (221, 339); (223, 324); (263, 333); (280, 337); (311, 332); (320, 325); (320, 353); (317, 373); (310, 376); (303, 383); (298, 383); (289, 385); (284, 380); (271, 378); (262, 374); (257, 373); (251, 373); (248, 369); (238, 366); (232, 361); (226, 358); (220, 355)
Posted by: Andrew Arnott at January 26, 2006 03:12 PM
I am writing my Greedy Algorithm and having issues with the GetPixelWeight method. It is almost always returning a negative number. What is a negative weight all about? I checked and the MaximumGradient value for the image was -1. What is up with negatives? If these are supposed to be this way, what are we defining as a lesser weight, a larger negative OR closer to zero? FYI: I was using the Parrot picture while this was occuring.
Posted by: Mike at January 28, 2006 04:35 PM
Mike,
See my comments on that at
http://vv.cs.byu.edu/cs312/archives/2006/01/a_small_error_u.html
It turns out that MaxGradientValue is not set correctly, but it shouldn't impact performance.
The simple greedy scissors are just lame, so don't expect them to work very well. You are probably already done with the algorithm and should move on.
"the other" mike.
Posted by: Mike Jones at January 28, 2006 06:28 PM
Dr. Jones,
What license is your source code for this scissors project distributed with? I blogged about my Dijkstra algorithm and some people are asking for the full program, but I wonder how you feel about it, as its original author. I'd like to respect your wishes with regard to your source code.
Posted by: Andrew Arnott at April 5, 2006 01:09 PM
Andrew,
I don't mind if you distribute with the following restrictions: 1. it can't be made publically downloadable because we are going to use it for future projects. If you want people to have to ask to get a copy, then that's fine. 2. When people start asking you for copies, just make sure they probably aren't in 312.
If you could set up a license to make those 2 restrictions transitive (that is, they apply to future users) then that'd be cool too.
Posted by: Mike Jones at April 6, 2006 01:23 PM
Post a comment
Thanks for signing in, . Now you can comment. (sign out)
(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)