> Maybe not only use multithreading for speedup, also need some
> algorithms to improve it.
> What about the speedup compare data?
Additionally, a fast implementation of box filtering is used and sub-sampling as recommended in https://arxiv.org/abs/1505.00996. Current implementation outperforms the existing bilateral filter in terms of speed and quality on my 8-core CPU. The speed depends significantly on the sub-sampling parameter. With sub = 4, for example, the filter works 110 FPS against 74 FPS for the bilateral one with the same spatial radius.

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".

Reply via email to