Convolution / Kernel / Feature map dimension
| Created | |
|---|---|
| Tags | CNN |
15) Describe how convolution works. What about if your inputs are grayscale vs RGB imagery? What determines the shape of the next layer? [src]
https://dev.to/sandeepbalachandran/machine-learning-convolution-with-color-images-2p41
- Convolution is a mathematical operation trying to learn the values of filter(s) using backprop, where we have an input I, and an argument, kernel K to produce an output that expresses how the shape of one is modified by another.
- Convolutional layer is core building block of CNN, it helps with feature detection.
- Kernel K is a set of learnable filters and is small spatially compared to the image but extends through the full depth of the input image.
- Dimension of the feature map as a function of the input image size(W), feature detector size(F), Stride(S) and Zero Padding on image(P) is (W−F+2P)/S+1
- No. of parameters = (Kernel size * Kernel size * Dimension )+1 = 28