Can someone explain these two to me?



Please avoid posting images from the quiz questions (and answers) on the Forum. Asking a question about the a quiz topic is fine, but please don’t discuss specific questions in detail.

Re: 1) Please review the Week 1 videos “Edge Detection Example” and “More Edge Detection”.
The key concept is that the symmetry in the filter determines what type of edge it will detect. The quiz asks you to expand your understanding of this concept beyond the two examples given in the lecture.

Re: 3) (more later…)

For the number of parameters in a CNN layer, please review the video “One Layer of a Convolutional Network” starting from time mark 5:22. For the example cited in your post, there are 3 input channels, and 128 output channels (because its the same as the number of filters). You also must include the bias unit that is added to each filter.

Thank you! Will keep in mind about the pictures from next time!