outline from gadient mask

  • 208 Replies
  • 30692 Views
*

yotamarker

  • Trusty Member
  • ********
  • Replicant
  • *
  • 611
    • battle programming
Re: outline from gadient mask
« Reply #195 on: October 14, 2017, 08:28:48 pm »
the a.eye would be getting a massive amount of images and it would be impossible to explain to it what each one is.
and now that I think about it NN would take too much run time.
what if I cluster together successive images that have a threshold of similar attributes ?

*

yotamarker

  • Trusty Member
  • ********
  • Replicant
  • *
  • 611
    • battle programming
Re: outline from gadient mask
« Reply #196 on: October 28, 2017, 12:16:12 pm »


1 2 3
4 5 6
7 8 9

x,y coordinates :

1,1 1,2 1,3
2,1 2,2 2,3
3,1 3,2 3,3

(0 to w div 3,0 to h div 3), _______(w div 3 to 2w div 3, same), (2w div 3 to w, same)
(0 to w div 3,h div 3 to 2h div 3), (w div 3 to 2w div 3, same), (2w div 3 to w, same)
(0 to w div 3, 2h div 3 to h), _____(w div 3 to 2w div 3, same), (2w div 3 to w, same)

(0 to w div 3,0 to h div 3), _______(w div 3 to 2w div 3, 0 to h div 3), _______(2w div 3 to w, 0 to h div 3)
(0 to w div 3,h div 3 to 2h div 3), (w div 3 to 2w div 3, h div 3 to 2h div 3), (2w div 3 to w, h div 3 to 2h div 3)
(0 to w div 3, 2h div 3 to h), _____(w div 3 to 2w div 3, 2h div 3 to h), ______(2w div 3 to w, 2h div 3 to h)



formula :

legend :
w : image width
h : image height
R : R*R = total desired shibers

(x-1)W div 3 to xw div 3, (y-1)h div 3 to yh div 3


(x-1)W div R to xw div R, (y-1)h div R to yh div R


div = \

*

Art

  • At the end of the game, the King and Pawn go into the same box.
  • Global Moderator
  • *********************
  • Deep Thought
  • *
  • 5106
Re: outline from gadient mask
« Reply #197 on: October 28, 2017, 01:16:01 pm »
What is a Shiber? Is not in any of my dictionaries. Thanks. Shiver perhaps?
In the world of AI, it's the thought that counts!

*

yotamarker

  • Trusty Member
  • ********
  • Replicant
  • *
  • 611
    • battle programming
Re: outline from gadient mask
« Reply #198 on: October 28, 2017, 01:33:02 pm »
What is a Shiber? Is not in any of my dictionaries. Thanks. Shiver perhaps?

I split the image into 9 images each of which I called a visual shiber.
those the process time is 9 times faster.

visual shiber = the most important area within the image to process at the moment

*

keghn

  • Trusty Member
  • *********
  • Terminator
  • *
  • 855
Re: outline from gadient mask
« Reply #199 on: November 03, 2017, 04:40:18 pm »

*

yotamarker

  • Trusty Member
  • ********
  • Replicant
  • *
  • 611
    • battle programming
Re: outline from gadient mask
« Reply #200 on: November 04, 2017, 01:30:55 am »

*

keghn

  • Trusty Member
  • *********
  • Terminator
  • *
  • 855
Re: outline from gadient mask
« Reply #201 on: November 16, 2017, 08:44:18 pm »

Computer Vision: Crash Course Computer Science #35: 


*

yotamarker

  • Trusty Member
  • ********
  • Replicant
  • *
  • 611
    • battle programming
Re: outline from gadient mask
« Reply #202 on: December 08, 2017, 12:21:28 pm »

*

Xman2000

  • Roomba
  • *
  • 2
Re: outline from gadient mask
« Reply #203 on: August 13, 2018, 02:46:34 pm »
Hy Yotamarker, i want colaborate working in your project, i am study about contour.
please share your Inventor source-code to download and the Yotamarker app source too.
you using only Harris operator?  cheers!

*

yotamarker

  • Trusty Member
  • ********
  • Replicant
  • *
  • 611
    • battle programming
Re: outline from gadient mask
« Reply #204 on: August 19, 2018, 03:09:03 pm »
Hy Yotamarker, i want colaborate working in your project, i am study about contour.
please share your Inventor source-code to download and the Yotamarker app source too.
you using only Harris operator?  cheers!



at any rate next stages with the A.I include : finishing the classifier for :
all objects detected, biggest object in region, outer outline, maybe some more attributes
speed up by taking out outline markers,
and using threading :
.net supports threading for loops but java doesn't (as far as I know) so I"ll have to make a decision there.

from there : the road splits for vb (connect to cam) and java : translate vb code + get code walkthroughs for getting cam images and
pixel attributes (RGB / HCL) and applying it to a service.

therefore I didn't have much time to work one it I was making these grimoire :
https://www.yotamarker.com/t155-android-mobile-app-dev-apk
and some more. BTW Xamarin is so dead  >:D  :knuppel2:

you got my PM ?

*

Xman2000

  • Roomba
  • *
  • 2
Re: outline from gadient mask
« Reply #205 on: August 21, 2018, 11:58:13 pm »
Yotamarker, you project of A.I. Girfriend is very cool !!!

i am know only VB6 (better), and VB variants language programming but i am start C# and will learning more languages.

i am working with same things you, same projects !!!

i am want make a APP software A.I. too but to put into a phisical, solid Robot, like a toy Robot, using Arduino.
i am still developping my projects in VB variants an after translate to other languages.
if you likes we can colaborate in both projects, my and your.

you need put your all Project Specs,Explanations, next Stages into a PDF to more clear you want and need, put into a WebPage too.
- SHAPE CONTOUR OF PERSONAGE AND graphics Programming
- BRAIN A.I. of PERSONAGE
- TASKS SKILLS WORKING OF PERSONAGE (talk, dance, controlling the things)
- NECESSARY RESSOURCE THINGS YOU NOT HAVE(  )
- PICTURES OF STAGES AND THINGS
- THINGS RESSOURCES YOU HAVE
- TIME OF DEVELOPPING EACH STAGE
- TIM OF CONCLUSION OF ALL PROJECT

why do not you use a Graphics Library wiht all Algorithms you need?

i think the problem of a Graphics Library is not have in VB but in Java or other .net have, and translate to VB is a lot of files, too many files.

You can switch a entirel Graphics Library like ACCORD.NET C#  to VB?
but i think you need a mutch more algorithms and switch a entirel Graphics Library maybe very useful to use in VB projects.

i know great algorithms and if you share your source-code of Inventor with me i can look how it works and put new algorithms to make this more powerfull.
you can understand the papers of Pseudocodes of Researchs of Algorithms in ResearchGate? you can switch the pseudocodes to VB?

i can share the algorithms with you and you returns the VB translated version with me?
i think you need works more now to better GRAPHICS stage, know and use more Algorithms, have Algorithms done to use if necessary and after next stage.
please, i need the source-code VB of INVENTOR if you want my colaboration. O0
I am not find the link to download the Inventor VB.net or Android APP, i not know how download?
I sent P.M now with more infos, look!

EDIT: My goal now is make a very good Image Contour app in VB to extract shape of image and create the shape of the Personage Character.
« Last Edit: August 22, 2018, 02:49:13 am by Xman2000 »

*

spydaz

  • Trusty Member
  • *******
  • Starship Trooper
  • *
  • 310
  • Developing Conversational AI (Natural Language/ML)
    • SpydazWeb
Re: outline from gadient mask
« Reply #206 on: August 23, 2018, 12:30:54 pm »
I have also worked on this Problem ; Also being a visual basic programmer (any) i have found there are some Library's for this in C#...
I found that some of the components for "CAMERA" are actually in the Universal samples

https://github.com/Microsoft/Windows-universal-samples/tree/master/Samples

Which are using the universal apps interface for programming to which i have not changed over to "YET"....

Although i was able to use the https://github.com/Microsoft/Windows-universal-samples/tree/master/Samples/BasicFaceDetection
& https://github.com/Microsoft/Windows-universal-samples/tree/master/Samples/BasicFaceTracking samples.

Actually microsoft do have a lot of library's available to which we usually go to ACCORD / NLTK / OPENCV for .... using the NuGet you can have the files downloaded into your app for compiling / referencing etc....

i personally use something similar to the openCV..... "haar cascades" ... for object/Face recognition...as well as a trainer application to create the cascade files... XML

Although i am now going towards Shape Recognition.... To black out the shapes or objects found.... then compiling a DB of simular shapes and using their properties as inputs for a NN which can give abstraction to the objects increasing the recognition of the object. Regions of the object can be used for detailed recognition ... ie FACE .... detect heads first generally then within object detect face..... allowing for deeper abstraxctions to be learned from characteristics such as sub objects in the prime object (HEAD/FACE/EYES) - EYE DISTANCE / EYE SHAPE / COLOR..... To get to these finite properties we need to dig into the picture to select just the components for matching with stored images/data......MAYBE

*

spydaz

  • Trusty Member
  • *******
  • Starship Trooper
  • *
  • 310
  • Developing Conversational AI (Natural Language/ML)
    • SpydazWeb
Re: outline from gadient mask
« Reply #207 on: August 23, 2018, 12:40:49 pm »
in understanding computer vision ;
We need to understand the generalisation of human vision...
How do we as humans perceive objects ...

we actually block most of the image out / by focusing on the target to be identified; drilling down further until the image is completely recognised ...


Something being the wrong colour or shape actually causes us delays in recognition .... relearning the new properties and adjusting out perception of the target being recognised........

And a lot more.......

*

yotamarker

  • Trusty Member
  • ********
  • Replicant
  • *
  • 611
    • battle programming
Re: outline from gadient mask
« Reply #208 on: August 26, 2018, 07:21:44 pm »
I just need the get image from camera repeatedly and fast programmatically on android studio.
GitHub is so difficult to understand, I need the step by step walkthrough. I read there are libraries and OpenCV and
GitHub projects but have no idea how to use them cause no walkthroughs. so I entered a full course about app dev.
and I simply neg the instructors till they cough up the walkthroughs I need.


BTW I could also use the get volume meter and amplitude walkthrough if someone has it

also the augmented reality walkthrough.

but mostly the camera.

also how to UL and DL android studio projects to that github. (poor instructors)  ::)

 


The many interfaces of computing
by Tyler (Robotics News)
November 14, 2018, 12:00:23 pm
Emotionally Responsive Avatars
by 8pla.net (AI News )
November 12, 2018, 10:00:17 pm
World's First AI News Anchor
by squarebear (AI News )
November 12, 2018, 08:22:59 am
Looking back at Project Athena
by Tyler (Robotics News)
November 11, 2018, 12:01:49 pm
A.I. Optometrist
by LOCKSUIT (AI News )
November 11, 2018, 07:06:36 am
Highlighting new research opportunities in civil and environmental engineering
by Tyler (Robotics News)
November 08, 2018, 12:00:38 pm
Why some Wikipedia disputes go unresolved
by Tyler (Robotics News)
November 07, 2018, 12:01:07 pm
To build new college, MIT seeks campus and alumni input
by Tyler (Robotics News)
November 05, 2018, 12:03:06 pm

Users Online

68 Guests, 2 Users
Users active in past 15 minutes:
sugumaton, LOCKSUIT
[Roomba]
[Trusty Member]

Most Online Today: 87. Most Online Ever: 208 (August 27, 2008, 09:36:30 am)

Articles