Lecture 4: Transforms (11)
charshou

If we know the camera properties as well as where they are in space, would we be able to convert from 2D to 3D coordinates, similar to how having two eyes provides us with a sense of depth?

brandonlouie

@charshou yes, I believe you can! I remember learning about this in EECS106A, and I think this discussion worksheet and its solution might be helpful in understanding how you can do so: https://ucb-ee106.github.io/eecs106a-fa23site/assets/disc/disc5_sols.pdf

(I would explain it myself, but I fear that my explanation might not be very helpful since that portion of the class feels like a bit of a blur to me ... :'))

AlsonC

@brandonlouie I really enjoyed this resource and thank you for it! I found a cool video that explored one of the topics mentioned in that document (convolutions) here: (https://www.youtube.com/watch?v=KuXjwB4LzSA&ab_channel=3Blue1Brown) which I found to be a great high level explanation of this topic.

AnikethPrasad

@AlsonC I watched through your video linked at found it very interesting. The use of kernels in image processing to blur/translate/sharpen is very interesting. It's pretty cool that we can edit photos in different ways simply by modifying our kernel.

You must be enrolled in the course to comment