From the course: AI Workshop: Hands-on with GANs with Deep Convolutional Networks

Unlock this course with a free trial

Join today to access over 24,900 courses taught by industry experts.

Applying convolutional and pooling layers

Applying convolutional and pooling layers

From the course: AI Workshop: Hands-on with GANs with Deep Convolutional Networks

Start my 1-month free trial Buy for my team

Applying convolutional and pooling layers

“

In this demo, we'll see how we can apply convolutional kernels of different types to the input image and look at the resulting image. We'll also see the result of applying a max pooling layer after we've applied the convolutional kernel. Every kernel that we apply will perform a different operation on the underlying image. For example, this very first kernel here is a sharpened kernel. The sharpened kernel has weights which serve to sharpen the underlying image. Sharpening is a combination of edge detection plus the original image. The sharpened kernel has a sum of one. Now how exactly this kernel sharpens the underlying image? That's a mathematical construct that we won't get into, but we'll focus on how to apply these convolutional filters to the input image and view the result. Observe that I've specified the kernel in four dimensions. This corresponds to the four dimensions in which the original image is expressed. If you remember the original image, that's a 4D tensor as well…

Contents