i want to implement this transformation formula on any image, give me code, links , tutorials related to it ?

how i implement this formula on such kind of images friends plz help

[chitresh]

[chitresh]

[chitresh]

this is the formula i am looking to implement and on such kind of images

Answer 1

[Image Analyst]

I don't know what x and y are. If they are coordinates then this t value does not depend on the image pixel intensities at all, which strikes me as strange.

[chitresh]

i think so this is x and y coordinates according to me so how we cone solve this if a and y where coordinated

[Image Analyst]

What is it? It sort of looks like some kind of image moment except for the fact that it doesn't depend on any image intensities. What happens if you just do it like it says, like :

n1 = sum(x .^ 2);
n2 = sum(y);
n3 = sum(x);
n4 = sum(x .* y); 
numerator = n1*n2+n3*n4;
d1 = n*n1;
d2 = n3^2;
denominator = d1-d2;
t = numerator / denominator;

[chitresh]

ok this i know too , still thanks main thing is how to get xdata and ydata, like i see some example where it is written [B,xdata,ydata] = imtransform(I,tform); where I is imread('Input_image'); but i dnt know that tform and whats use.... or there is some other way to get xdata and ydata

[Image Analyst]

You're not answering my questions, and I don't know how to answer yours, especially when vital details are released in little bits and pieces over several posts. I still have no idea what you want to do, where you got that equation, why you used imtransform(), etc.

[chitresh]

i just want to know xdata and ydata for any random image

[Image Analyst]

It looks like it gets returned from imtransform().

[chitresh]

dnt know i am looking for some tutorials too to get xdata and ydata values

[Image Analyst]

The example on "Projective Transformation" in the help does that.

[chitresh]

here my x and y is the 2 dimension vector point can you tell how to find these vector points and here i is the user and n is the minutiae points, so guys now hope i am clear , i got all stuff on a paper...

[Image Analyst]

I hope my code helped you. It does what you want, even though I don't understand what you want represents, or why you want it. Does this solve everything for you now?

Answer 2

[Roger Stafford]

Chitresh, I can tell you what kind of mathematical problem your expression is a solution to. Let us have n x and y pairs

 x = [x1,x2,x3,...,xn]
 y = [y1,y2,y3,...,yn]

and suppose we want to find the line of regression through them - that is, to find constants a and b such that

 sum((a*x+b-y).^2)

is a minimum. This is a least squares problem and the solution for the constant b is

b = (sum(x.^2)*sum(y)-sum(x)*sum(x.*y))/(n*sum(x.^2)-sum(x)^2)

which is precisely your expression.

I can only guess how this might apply to the fingerprint image you display. If the xi and yi pairs refer to pixel coordinates of the black areas, and if the coordinate origin is at the lower left corner of the image, then b would be the vertical distance above the bottom at the left side where the line of regression crosses. On the other hand if the origin is located in the center of the image, b would be the vertical distance above or below this origin where the regression line crosses a vertical line through that origin. I would have thought the expression for a would have at least as much significance in the image as with b, since it would give the slope of that line of regression.

There are probably several other possible meanings that the x and y pairs might possess with regard to the image. I think that is a puzzle you will have to crack. Note that when you do solve your puzzle, the above expression shows you how to implement it in matlab in one line.

[chitresh]

here my x and y is the 2 dimension vector point can you tell how to find these vector points and here i is the user and n is the minutiae points, so guys now hope i am clear , i got all stuff on a paper...

[Image Analyst]

OK, so x and y are coordinates of minutiae. But what does t in your formula represent? If it has something to do with fitting a line, how does that make sense? The minutiae don't lie in a line.

[chitresh]

this is some kind of transformation formulae so do you have any idea or links or tutorials related to it

[Image Analyst]

Sorry, no.

Answer 3

[msahar]

I am also working on matlab image transformation, affine, rotations, point clouds, normal unifications. If you tell me in detail your problem, may be i can help you.

[chitresh]

10 degress between 0 and 180 degrees on each N x N square in the image. The important parameters for a gabor filter are its frequency, f, it’s standard deviations in both the x and y directions and its directional strength, gamma (multiplied by the y’2 term). The frequency, f, represents the inverse of the average number of pixels between ridge centers.

[chitresh]

i want to implement this i already implemented normalization formulae now i like to implement this