[jitter] Re: shift register
Wesley Smith
wesley.hoke at gmail.com
Tue Apr 1 14:24:23 MDT 2008
- Previous message: [jitter] Re: shift register
- Next message: [jitter] Re: depth message wouldn't resolve trails ?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
On Tue, Apr 1, 2008 at 2:13 PM, Robert Ramirez <rob at robtherich.org> wrote: > > yeah, that's a beauty. > > so, just for my own clarification, when the matrix is sent through jit.convolve, it's making each pixel be the exact value of the pixel immediately to it's left? > Based on the kernel in AB's patch, it does exactly this because there's a gain of 1 in the left-of-center cell of the kernel. Typically convolution is a somewhat heavyweight operation, but jit.convolve has some optimizations for certain kernels where many of the weights are 0. weds
- Previous message: [jitter] Re: shift register
- Next message: [jitter] Re: depth message wouldn't resolve trails ?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
