Senior Thesis

A place to discuss the science of computers and programs, from algorithms to computability.

Formal proofs preferred.

Moderators: phlip, Moderators General, Prelates

freedyvorhees
Posts: 6
Joined: Wed Jun 08, 2011 8:40 am UTC

Senior Thesis

Postby freedyvorhees » Wed Jun 08, 2011 11:44 pm UTC

Could you suggest a topic for a computer science thesis. My adviser specializes in image processing topics. I suggested a low res optical-character-recognition in real-time. But he rejected it. He said its to hard. Do you have some ideas?

kmatzen
Posts: 214
Joined: Thu Nov 15, 2007 2:55 pm UTC
Location: Ithaca, NY

Re: Senior Thesis

Postby kmatzen » Thu Jun 09, 2011 5:21 am UTC

Make a "cartooning" filter and put it on a cell phone.

freedyvorhees
Posts: 6
Joined: Wed Jun 08, 2011 8:40 am UTC

Re: Senior Thesis

Postby freedyvorhees » Thu Jun 09, 2011 8:02 am UTC

How "cartoon"? Do you know any sites where I can read some literature?

Moose Hole
Posts: 398
Joined: Fri Jul 09, 2010 1:34 pm UTC

Re: Senior Thesis

Postby Moose Hole » Thu Jun 09, 2011 1:27 pm UTC

How about trying some unlimited detail technology?

User avatar
WarDaft
Posts: 1583
Joined: Thu Jul 30, 2009 3:16 pm UTC

Re: Senior Thesis

Postby WarDaft » Thu Jun 09, 2011 3:07 pm UTC

Moose Hole wrote:How about trying some unlimited detail technology?

That stuff's just a crock.


How about writing and training an ANN which does OCR? They're quite well suited to that.
All Shadow priest spells that deal Fire damage now appear green.
Big freaky cereal boxes of death.

kmatzen
Posts: 214
Joined: Thu Nov 15, 2007 2:55 pm UTC
Location: Ithaca, NY

Re: Senior Thesis

Postby kmatzen » Thu Jun 09, 2011 3:29 pm UTC

freedyvorhees wrote:How "cartoon"? Do you know any sites where I can read some literature?

http://www.cs.northwestern.edu/~holger/ ... action.pdf

This is easy to implement using OpenGL ES 2.0 on a modern cell phone.

freedyvorhees
Posts: 6
Joined: Wed Jun 08, 2011 8:40 am UTC

Re: Senior Thesis

Postby freedyvorhees » Tue Jun 14, 2011 12:07 am UTC

How about writing and training an ANN which does OCR? They're quite well suited to that.


This is interesting. Do you have a literature?

freedyvorhees
Posts: 6
Joined: Wed Jun 08, 2011 8:40 am UTC

Re: Senior Thesis

Postby freedyvorhees » Wed Jun 15, 2011 4:21 am UTC

My adviser suggested this http://www.geocities.jp/igarashi_lab/papers/holly_npar09.pdf, any thoughts?

Parsifal
Posts: 113
Joined: Thu Feb 28, 2008 1:35 am UTC

Re: Senior Thesis

Postby Parsifal » Wed Jun 15, 2011 11:05 am UTC

I thought the point of a thesis was to demonstrate your capability to conduct and publish original research? How would your teacher know if it is too hard? "Too hard" means too hard for your teacher to understand and/or grade, but I digress.
The stencil app would indeed be easy to implement in OpenGL, AWT or even HTML 5 & Javascript. If the assignment is to develop your own algorithm for finding islands (isolated positive regions) in a stencil template and optimally connect them, sounds like a decent idea to me.

Parsifal
Posts: 113
Joined: Thu Feb 28, 2008 1:35 am UTC

Re: Senior Thesis

Postby Parsifal » Wed Jun 15, 2011 11:07 am UTC

Moose Hole wrote:How about trying some unlimited detail technology?


Excellent idea, sir. Please cite a peer-reviewed article so this student knows what to implement.

Thought so.

freedyvorhees
Posts: 6
Joined: Wed Jun 08, 2011 8:40 am UTC

Re: Senior Thesis

Postby freedyvorhees » Fri Jun 17, 2011 11:19 am UTC

Parsifal wrote:I thought the point of a thesis was to demonstrate your capability to conduct and publish original research? How would your teacher know if it is too hard? "Too hard" means too hard for your teacher to understand and/or grade, but I digress.


Well in our college, to fulfill a thesis for a bachelor's degree at least you have to recreate an existing original study with your own results and insights, and if possible your own algorithm. And maybe what he meant was that its too hard for me. We were given 5 months. And I still have 7 other subjects. And I don't have the slightest idea how to do an OCR. But I find it interesting.

Parsifal wrote:The stencil app would indeed be easy to implement in OpenGL, AWT or even HTML 5 & Javascript. If the assignment is to develop your own algorithm for finding islands (isolated positive regions) in a stencil template and optimally connect them, sounds like a decent idea to me.


That's exactly what he wants me to do. He doesn't prioritize a GUI for the project, but its encourage. I think he suggested OpenCV for the development. And if you think its a decent idea, I think I'll give it a shot.

Do you have other ideas?

Parsifal
Posts: 113
Joined: Thu Feb 28, 2008 1:35 am UTC

Re: Senior Thesis

Postby Parsifal » Sat Jun 18, 2011 12:36 pm UTC

There's nothing wrong with the idea your professor suggested, and you could probably spend a month or more making a robust application that implements the algorithm as specified in the paper you cited. Then, you can substitute the stencil filter implementation for one of your own design and compare the results (speed, stencil quality, performance vs pathological cases). It would be a nice touch to find a case that seems to break the published algorithm and try to come up with a way to handle it.

Also, working on the project your professor suggested means that he wold probably be more likely to provide guidance if you get stuck.

Also, if you want to use opencv and don't currently know any of the supported languages, that would be a learning opportunity.

freedyvorhees
Posts: 6
Joined: Wed Jun 08, 2011 8:40 am UTC

Re: Senior Thesis

Postby freedyvorhees » Mon Jun 20, 2011 9:11 am UTC

Thank you sir. Tomorrow will be the beginning of many a hard days night.


Return to “Computer Science”

Who is online

Users browsing this forum: No registered users and 6 guests