DLS Course 4 Week 4 Quiz

Course Q&A Deep Learning Specialization Convolutional Neural Networks
1

Hi,

One of the questions in the C4 W4 Quiz asks about the number of output units of the landmarks.

"Question:

When building a neural network that inputs a picture of a person’s face and outputs N landmarks on the face (assume that the input image contains exactly one face), we need two coordinates for each landmark, thus we need 2N output units. True/False?"

The answer is marked as True. However, doesn’t the output need one more unit telling the probability of whether there is a face or not? At least that’s how I understood Andrew’s explanation (at 2:37 of the recording). Thus, 2N+1. His note says 129 units output for 64 landmarks. Can someone explain why the Quiz says 2N is correct?

Thanks!

2

The quiz says 2N because the scope of the question is the number of units. It does not include the detection output.

3

I have similar question in a different way.

image
image808×248 12.9 KB

I think the answer to shape of y estimate is (1, 2N), but got wrong.
Then I’m confused what the shape of y should be? Is it in the options?

4

(2N,1) or (1,2N) is determined by how you design the NN. Either one could be true.

But the last statement is always true.

5

The thing is after I tried (1,2N), it said wrong, giving the reason " y hat is column vector, not row vector". Then I tried the last option, but still wrong. That’s why I am confused.

Thank you for the confirmation. At lease I have an answer.