Hello Everyone!
As i am doing my master thesis under the topic "Data driven depth estimation from a monocular image using deep neural network", what concerns me more is the output of my neural network.
My input will be an image of what soever size and my CNN network will learn some features out of it. What I don't understand is what will be my output and how can i get a color visualization of that depth data from neural network?
I have read several papers where it says our model predicts depth. What I don't understand is how are they getting a color visualization from an output of a neural network?
Thankyou , any kind of help will be appreciated!
I think that It's possible that they are applying the color visualization once the depth estimation is already done. Which actually means combining the input image with the depth data estimated from the CNN.
@MohammedBenmahdjoub can you please elaborate a little on your last sentence.Thankyou!
I meant that for each depth point (x,y,z) they associate with it (r,g,b) value that they have from the initial image. (it's a mapping : (x,y,z) gives you the color (r,g,b)). So basically what I mean is that I think that the color visualization has nothing to do with the CNN.
(PS: that's an answer for the second point about the other papers. I don't have an answer for your first question). Good luck Vishesh Breja
- Badges
- Science topic
- Similar topics
- Papers