Guys and "grills", I finishing my degree in Comp Sci

Come on guys, I'm kinda desperate.

>>8750938
a faster square root.

>>8750922
>computer graphics
Optical flow

The idea is you take 2 images separated by a short amount of time (i.e. take a picture, then take a 1/2 step forward, and take another picture)

then look at the differences between them, the movement of the blocks of pixels, etc. and come up with an algorithm to determine which objects are close and which ones are far.

good shit. look into it

if you come up with a sweet enough algorithm, you'll win a nobel and you'll get hired by google or amazon or someone for 100k+ or win a nsf fellowship and get paid to go to berkeley

>>8750922
p=np

>>8750995
>8750995
Thx, I'll look into it

Build a 3D scanner. The only hardware is a rotating table driven by a motor with a rotary encoder (so you know what angle the object is as), a laser, and a camera. Beyond that it's image recognition, edge detection, and CompSci math.

>>8750997
too difficult, I already tried modeling a CSP that is classified as NP-Hard and creating a polynomial algorithm to it in order to proof this, but I failed too many times.

Also, I lack the capabilities for elaborate a proof that P!=NP.

Thx for the suggestion anyway.

>>8751073

>>8751064
It's a good idea, but I don't think I can make a senior thesis out of this project. It's quite simple and the construction of a 3D scanner is already well documented.

I'm studying volumetric rendering tough.

>>8751073
Yes, he'll probably use some recursive algorithms

3d printing. If you see the option, you catch my drift and you WILL have a good one

>>8751073
Your post caused me a stack overflow.
I need an aspirin.

>>8751073
It's odd how that doesn't register as it's own reply

>>8750922
deep learning as it apples to malware analysis

Before this thread go 404 I will say that I'm leaning towards Optical Flow

>>8750995
First thing in my mind when I read this was autopilot for a missile.
>>8751208
If you pick this subject and do it well, you may not be able to post it here, export restrictions on weapons.
But that will get you a job.

>>8751260
This would probably fit in a Computer Engineering senior thesis, But I think an embedded system aided to computer vision would be a good topic

>>8751073
but how??

>>8751208
image processing is for faggots who gag on massive african dicks

File: Alfred_Binet.jpg (721KB, 2188x2812px) Image search: [Google]

721KB, 2188x2812px

>>8751073
>
fucking what?

I am into the same problem, OP.
I will probably try to work with Comp Linguistics, or natural computing

>>8751073
Someone with decent technical knowledge tell me what this is and how he did it?

>>8750922

>>8751061

protip: you're supposed to use hamming_dist8() on the R,G, and B of each image

>>8751533
that's one potential algorithm.

>>8751260
>>8751208
it's actually a hot topic right now. do a thorough prior art search. im sure your college gives you access to tons of journals and publications. coming up with a working, efficient implementation of one of these algorithms would be enough for a senior thesis most likely.

>>8751434
Interesting, I studied a bit too, but it envolves a lot of functional programming, something that I'm really not a fan.

>>8751466
On a slow board like /sci/ it's possible to figure out the post number you'll get when you post before it's actually oisted. From there it's a simple matter of including that number into the post as you would a normal reply.

Fuck your interests, solve a problem in complexity theory, a lot of low hanging fruit out there if you are dedicated enough.