Training classifer Manual..

[ramisetti]

hello all

i need a small help in my project i need to do object classification based on their color and size. As i know that by using "color classification training" i can train my classifier with color but  i shouldnt use it i need to create a program my own training phase by using labview vi´s.

is it possible to do it by using labview classification vi's like imaq write classifer imaq read classifier imaq add sample etc...

i am so confused how to use them....

[TiTou]

There is an example that ships with LabVIEW + VDM :

C:\Program Files\National Instruments\LabVIEW 2014\examples\Vision\Classification\Color Classification.vi

Take a look at it 😉

[Hatef.fouladi]

there is some thing that you have to explain before start to coding 
1) what is Priority for classify color or size ?
2) what is your means from size? area or biggest diameter of objects or surrounding ?
3) do you have to use labview classify method to solve this problem?

[ramisetti]

hi,

by using the reference of the color classification training in vision assitance i tried to design my own classifer training by using imaq create color classifer ,imaq color classifier options imaq add classifier sample, imaq train nearest  neighbour imaq write classifer.

but i am not able to sample in the vision assistance i took a roi and just keep on adding sample but here i am not able to understand how to do it.

i ma attaching the screen shots and the vi please help me with your suggestions.

[ramisetti]

hello titou

i saw that example but that example is based on the .clf files created on the color classification interface but in my condition i need to create my own training interface in labview.

[Hatef.fouladi]

you do not need to use labview classify it is easy code to make for your goal but fist you have to answer the question for finding best code

[ramisetti]

hello Hatef,

Based on the color i need to classfiy the object. i need to use labview classify method to classify the object but before classification i need to train classifer so used vi present in labview 2013 to code for the training classifer as my idea i designed as follows:

imaq create color classifier vi creates classifier session and imaq color classifer option vi configures my classifier session with their color options low,medium or high intensities.

usng imaq add samples vi add the number of color samples to the defines color class name such that when i keep on object infront of my camera it has to auto classify the color of the object. But i am have no idea how to add samples to the class usin imaq add sample vi.

i should not use vision color classification training to train my classifier. so please help me with your valuable suggetions to overcome my problem.

[Hatef.fouladi]

ok before start to work you have to make your project simple as much as possible
for example use such code will be solve your problem
step 1
use Imaq article analysis to sort particle size and find appropriate ROI
step 2
use Imaq color match vi with ROI that you find in step one
select your sensitivity of image
and now you can classify your object up to 52 different row color area
isn't enough for you ?
I think it will be work just try it

[ramisetti]

hello hatef,

As i am very new to labview i am not able to understand it properly can you please refer any example regarding it so that i can understand it by reading it.  actually the vi i designed is based on the ieee paper i read  in which they explained one way of designing the classifer trainer. can you please look at it and tell me is it possible to do it that way.

i searched for my example s to understand this concept i really didnt found anything related to this topic. 

[Hatef.fouladi]

ok so let do it by one by and step by step
for start please try to work with particle analysis vi
there is some example in LV for that
open them and see how it work
particle analysis vi
Returns the number of particles detected in a binary image and a 2D array of requested measurements about the particle
the example is
LabVIEW\examples\Vision\2. Functions\Binary Analysis\Particle Analysis Example.vi