Gradient feature extraction matlab code

Can anyone explain how to make Gradient feature vector? For example, if we take Sobel edge detection operator, then how to make feature vector?

Answer 1

[Image Analyst]

There is edge() and imgradient() and imgradientxy(). So, once you have your edge image, what values from it do you want to stick into your feature vector? If you don't know, then we certainly don't know. You could stick virtually any number you want into a vector.

[Shivang Patel]

So, That mean value of Gmag , and Gdir are the feature vector...

Thank you..

[Image Analyst]

Could be, if that's what you want to be in your feature vector. Again, it can be whatever you define it to be.

Answer 2

[Diksha Kurchaniya]

would anyone can tell me how i can extract feature from sobel oprator. means i have get an image after applying the sobel operator to an image but how can i convert to feature vector to it

[Image Analyst]

You're asking the question backwards.

You don't say "let me apply some arbitrary algorithm and see what features I can get from it."

You need to say, "these are the features I need to identify my object, now what algorithm can get me those features?"

Answer 3

[Saeid Abbaasi]

for image name I, use the following code:

xHOG = extractHOGFeatures(I);

its a gradient based feature called Histogram of Gradient