Project

General

Profile

Feature #8861

Adaptive Manifolds for Real-Time High-Dimensional Filtering

Added by Reuben Martin about 7 years ago. Updated almost 6 years ago.

Status:
Closed: won't fix
Priority:
Low
Assignee:
-
Category:
Darkroom
Target version:
-
Start date:
08/16/2012
Due date:
% Done:

0%

Estimated time:
Affected Version:
System:
bitness:
64-bit
hardware architecture:
amd64/x86

Description

If you haven't seen this SIGGRAPH paper yet, you will probably find it very interesting. Has potential to both speed up and improve several of your filters when processing the filter stack. (or just provide means for creating completely new filters) Includes some example code.

http://inf.ufrgs.br/~eslgastal/AdaptiveManifolds/

History

#1 Updated by Johannes Hanika about 7 years ago

  • Priority changed from Medium to Low

it includes matlab code, which is not useful. plus, using 15 manifolds is exactly the same storage as the bilateral grid for 3d in practise, so i'm not immediately convinced it will improve things for us. also their non-local means results seem to employ pca for reduction of dimensionality and seem to be less optimal than specialized implementations.

that being said i'd like to see a real comparison and found the paper very interesting and exciting to read. so if anyone implements it, let me know how it goes (i won't code it).

#2 Updated by Reuben Martin about 7 years ago

The devs for both the Blender Cycles and Luxrender rendering engines were talking about using it to reduce noise and hot pixels since that was where it seemed to perform so well in the example images. Don't know if anything will come of it.

#3 Updated by Tobias Ellinghaus almost 7 years ago

  • % Done changed from 0 to 20
  • Status changed from New to Incomplete

#4 Updated by Tobias Ellinghaus almost 6 years ago

  • % Done changed from 20 to 0
  • Status changed from Incomplete to Closed: won't fix

One year without any news, I am closing here. Feel free to bump if anything new happens about it.

Also available in: Atom PDF

Go to top