Course 4. Why so many filters in CNN

[gadolinis]

Hey,

I have a general question about Convolutional NN. We do not specify what kind of filters we use in CNN, right? Are these filters set different automatically? I mean filters for vertical, horizontal edges and many other types. Otherwise why we should use so many filters in CNN?

Thank you in advance!
Best regards,
Gediminas

[TMosh]

The actions of the filters is what is automatically learned. They’re not necessarily vertical and horizontal edges, it depends on what characteristics in a training set might give the lowest cost.

We need lots of filters in order to learn complex abstractions about the data set.

[ai_curious]

@gadolinis there are several other related threads that you can find with search that might add to your understanding of this concept. Here some examples:

Are filter parameters all learned or specified?

TMosh and AI_curious Thx for both responses. I totally get that filter parameter values are learned - that is the foundation of deep learning. Based on the “known” filters, such as edge detection, my initial impression was that somehow the deep learning also figured out those same known filters. But now I am realizing that the learned filters have no direct correlation with those “known” filters at all. Those value work in that model, and that is it. Bottom line - understanding known filters …

Where are edge detectors?

In week1 of the Deep Learning course 4, the edge detectors are explained. When implement the code in assignment, we just give how many filters and kernel size we will assign in each layer. This look like, to me, the values (weights) in the kernel are not specific for edge filtering. Why the edges filters are not used? or when do I need to enforce edge filters? Best

Try searching on Edge detector or Edge filter

https://community.deeplearning.ai/search?q=Edge%20filter