Ask Your Question
-3

3d data recognition

asked 2017-04-28 06:51:37 -0600

Fredreload gravatar image

updated 2017-04-29 06:17:09 -0600

I want to know the algorithm that I can apply to 3D data and I would be able to recognize the objects inside. For instance if I use a laser scanner for my room to generate a 3D data chunk and the object is separated either by its reflection or edges. So for instance if I found an object with many pages I would recognize it as a book, if I find a chair object I would recognize it as a chair and so on. Is there an algorithm for me to do this? It would be used for DNA nucleotide identification(ACGT).

It is even better if it is an Opencv program I can work on. Thanks

image description

edit retag flag offensive close merge delete

Comments

1

Can you explain relationship between "DNA nucleotide identification(ACGT)" and chair ?

LBerger gravatar imageLBerger ( 2017-04-28 09:51:56 -0600 )edit

Well, my original intention is to identify the DNA nucleotide based on the protein structure. Any hint would be helpful

Fredreload gravatar imageFredreload ( 2017-04-28 09:54:16 -0600 )edit

Can you give a reference of 3d DNA data ?

LBerger gravatar imageLBerger ( 2017-04-28 09:58:06 -0600 )edit
1

Something from the Wikipedia https://en.wikipedia.org/wiki/Nucleobase (page), check out the pictures on the right for adenine, guanine, etc. Well, the real in vitro scan might not look exactly as is, I am speaking about it theoretically. But it is alright to prepare if someone ever makes a DNA structure scan.

Fredreload gravatar imageFredreload ( 2017-04-28 10:17:42 -0600 )edit

Opencv is mainly used for 2D image processing. You can read this tutorials may be it can help you

LBerger gravatar imageLBerger ( 2017-04-28 10:34:31 -0600 )edit

Hmm, what software would you recommend for 3D data recognition?

fredreload gravatar imagefredreload ( 2017-04-28 12:35:56 -0600 )edit

sorry no experience => google

LBerger gravatar imageLBerger ( 2017-04-28 13:02:12 -0600 )edit

How about an algorithm?

fredreload gravatar imagefredreload ( 2017-04-28 13:58:29 -0600 )edit
LBerger gravatar imageLBerger ( 2017-04-28 14:03:00 -0600 )edit

how do you even plan to get 3d data from your nucleotides ? you probably can't scan them with a laser ?

berak gravatar imageberak ( 2017-04-29 01:25:11 -0600 )edit

Berak: Long time no see. I have a theoretical idea presented here. Be sure to check the comments I posted. You are on the right track with a laser, after you laser scan the object it would reflect the electrons and you magnify it with a lens like this. It is theoretical because lens magnification seems like magic to me lol, but hey no ones ever tried to magnify the laser scanning technology with a lens, it's used in an electron microscope though Leberge: I got it to work, lots of version changes there and I had to create a swapfile but it looks cool, got other point cloud files .pcd I can test out?

Fredreload gravatar imageFredreload ( 2017-04-29 05:04:36 -0600 )edit

Well, I dunno mate, I tried to email MuAnalysis, the one with an electron laser scanner based on the intensity, but they did not reply to me. I'll hope they create a better design with my method

Fredreload gravatar imageFredreload ( 2017-04-29 06:19:54 -0600 )edit

Anyway, no hard feelings :D

Fredreload gravatar imageFredreload ( 2017-04-30 11:28:49 -0600 )edit

1 answer

Sort by ยป oldest newest most voted
0

answered 2017-04-28 19:45:07 -0600

Tetragramm gravatar image

Try OctNet. It's probably not suitable for your situation due to the way it quantizes data, but at least read the paper and check other citations.

https://github.com/griegler/octnet

edit flag offensive delete link more

Comments

Is there a clip showing how the project runs? I am having trouble compiling cuda, it looks promising

Fredreload gravatar imageFredreload ( 2017-04-29 10:24:17 -0600 )edit

Question Tools

1 follower

Stats

Asked: 2017-04-28 06:51:37 -0600

Seen: 528 times

Last updated: Apr 29 '17