Thanks for this video. I’m glad you touched on the cross-correlation vs. convolution operation. When I took computer vision last year, I was so confused why we were flipping kernels when we were convolving images despite having messed around with CNNs prior.