hi i making a multituouch surface with pd, i want be able to filter the low and middle bands in the video from my web cam. im interesting in just letting the high images pass... is there something like a high pass filter for imagens?
thanks
Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
Hi,
interesting project, how are you doing it? Back-projection and a IR-cam? I am not an image processing whiz, but you could try to subtract your image from a lowpass-filtered copy of the same image. Should give a highpass then. But i am sure there are more clever ways of doing it. Basically a HP can be done using a delay and a subtraction.
regards, PP
hi i making a multituouch surface with pd, i want be able to filter the low and middle bands in the video from my web cam. im interesting in just letting the high images pass... is there something like a high pass filter for imagens?
thanks
Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Do you want to out pixels that are constant in time, or are you looking for a 2-dimentional spacial highpass?
-spencer
On 11/5/07, punchik punchik punchikk@yahoo.com wrote:
hi i making a multituouch surface with pd, i want be able to filter the low and middle bands in the video from my web cam. im interesting in just letting the high images pass... is there something like a high pass filter for imagens?
thanks
Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
punchik punchik wrote:
hi i making a multituouch surface with pd, i want be able to filter the low and middle bands in the video from my web cam. im interesting in just letting the high images pass... is there something like a high pass filter for imagens?
in Gem: [pix_convolve] does spatial convolution with a square kernel. fill it with the right coefficients to produce a spatial high-pass. [pix_tIIR] does temporal filterin using an IIR-filter of settable length. fill with the right coefficients to produce a temporal high-pass. unfortunately the taps in the filter store the images only in the usual precision (e.g. RGBA32), which might yield weird results
in pdp and Gridflow there are surely solutions (which i don't know; just go through the manual), for spatial filtering see [pdp_convolve] (pdp) and [#convolve] (Gridflow).
pdp will work with higher precission (16bit instead of 8bit per channel), and GridFlow will even allow you to use floating point.
mfa.cae. IOhannes