PDA

View Full Version : state-of-the-art volume rendering algorithm



witand
02-13-2017, 06:33 AM
Hello,

I recently started to look into opencl as I plan to use it for a project that requires volume rendering. I read about different approaches for volume rendering and it seems that raycasting is the best but also most expensive.

It seems like the algorithm itself is not that complicated (not that many lines of code). Therefore, is there a big difference (in performance and visual quality) between different implementations?

Do you know a source for a state-of-the-art implementation that allows me to have a look at the source code and play around?

Thank you!

Dark Photon
02-13-2017, 06:07 PM
I read about different approaches for volume rendering and it seems that raycasting is the best but also most expensive. It seems like the algorithm itself is not that complicated (not that many lines of code). Therefore, is there a big difference (in performance and visual quality) between different implementations?

There are a lot of volume rendering techniques. Before you go surveying them, it's probably worth it to first determine your requirements. What's your input data look like? What effect/look are you trying to achieve? What's your performance target (min/max time per frame)? And if you're aiming for a GPU implementation, what GPU vendors and GPUs do you intend to support?


I recently started to look into opencl as I plan to use it for a project that requires volume rendering

Let's find out your requirements first. But there are some pros and cons for choosing OpenCL.