Convolution filter optimization (Feature #2992)
| Status: | Done | Start date: | 2013-04-26 |
|---|---|---|---|
| Priority: | Normal | Due date: | |
| Assignee: |  Vadim Pisarevsky |
% Done: | 0% |
| Category: | core | ||
| Target version: | - | ||
| Difficulty: | Pull request: |
Description
I've stumbled upon the following issue : I want to apply a fixed filter (box filter for example) on a 8U image so I shouldn't be using float should I?
Yet the code has been generalized and is using doubles to scale results. I'd love to have an option to tell filter to keep using integers
Associated revisions
#2992: update after review
Merge pull request #2992 from snikulov:upmerge_jpeg_tweak
History
Updated by Anna Kogan almost 12 years ago
Hello Sebastien,
Thank you for reporting the feature. If you could fix the it on your side, a pull request in our GitHub repo would be highly appreciated!
Updated by Daniil Osokin about 11 years ago
- Tracker changed from Bug to Feature
Updated by Vadim Pisarevsky almost 10 years ago
in 3.0-pre OpenCV uses IPP, which is hardly optimized, including boxFilter support
- Status changed from Open to Done