You explained the most direct difference between those two. The more interesting observation is that convolution is associative, while cross-correlation is not. That's why, for example in image processing, convolution is more commonly used.
This is not very useful. I watched to understand the intuition behind using these two. But, unfortunately, you only give the information about how to apply the formula on an example. You need to give the intuition behind using these methods (particularly in computer vision) to better understand their differences.
I think I get it now, so i and j are indices in our kernel, and when we do a summation over the expression using cross correlation / adding the summation counters elements are placed one to one with the original kernel, whereas when you introduce the subtraction of the summation using convolution, the elements are placed in reverse order going from left to right and up to down. Is that right, at least for this example?
![Convolution vs. Correlation in Signal Processing and Deep Learning [DSP #10]](http://i.ytimg.com/vi/R7cn9b7BNyk/mqdefault.jpg)
You explained the most direct difference between those two. The more interesting observation is that convolution is associative, while cross-correlation is not. That's why, for example in image processing, convolution is more commonly used.
Thank you!
This is not very useful. I watched to understand the intuition behind using these two. But, unfortunately, you only give the information about how to apply the formula on an example. You need to give the intuition behind using these methods (particularly in computer vision) to better understand their differences.
I think I get it now, so i and j are indices in our kernel, and when we do a summation over the expression using cross correlation / adding the summation counters elements are placed one to one with the original kernel, whereas when you introduce the subtraction of the summation using convolution, the elements are placed in reverse order going from left to right and up to down. Is that right, at least for this example?
what is the cause for this so-called "flipping"?
Watch udacity's computer vision filtering and linearity lesson. You will understand why.
I'm bit confused about the Symbols used in this video, Please Help!!!
I have added a simple correlation filter matlab function. Hope it helps..!
gist.github.com/0fb47ed53d4806f417b5f0e1f9922ca0.git
I am taking a deep learning course and this was very useful.
Thank you for this video!
Correlation is Similarity / Convolution is filter
this is my understanding
found nice article
towardsdatascience.com/a-comprehensive-introduction-to-different-types-of-convolutions-in-deep-learning-669281e58215
@@横川俊介-x4f thank you for sharing 👍
This is one of the worse explanations ive heard in my life
Are you crazy or what?.Explain something please
thanks saved my ass