# How do convolutional neural networks deal with many filters during convolution?

I am unsure of how convolutional neural networks treat several filters. Many of the examples I have seen only have filter at a time, and that is intuitive for me. Look at the nice visual tutorial here: https://www.jeremyjordan.me/convolutional-neural-networks/. If I look at the following visualization, I am okay with how they get from the input to the feature maps in the lower level concepts. However, I am unsure about how they get from the feature maps in the lower level concepts to the feature map in the higher level concept, and that largely stems from my confusion about how are multiple filters incorporated?

• A Convolutional layer simultaneously applies multiple filters to the input, making it capable of detecting multiple features anywhere in its inputs. So as you can understand easily, in the second block, things are run simultaneously. The output of this block will be $3 \times 3 \times 2$, Here the last dimension 2 represents number of feature maps generated. In the third block, two square detector filters have been applied and simultaneously computed the feature maps and summed the individual elements of the matrices. – ARAT Jan 10 at 21:28