Previous | Next --- Slide 23 of 38
Back to Lecture Thumbnails

I imagine 2D Directions could be done easiest by uniform sampling in theta, but I see that sampling directions in higher dimensions would be very intuitive by using rejection sampling?


Rejection sampling works, but there are also mappings for sampling directions directly using uniform random numbers (like the one for disk sampling). In a few lectures we'll come back to those (and the question of why one might prefer those to rejection sampling..)