[PD] [pix_blob] output coordinates

Peter P. peterparker at fastmail.com
Mon Nov 20 14:50:22 CET 2017


* Max <abonnements at revolwear.com> [2017-11-20 12:38]:
> On 2017년 11월 19일 23:06, Peter P. wrote:
> > (replying to myself below)
> > 
> > * Peter P. <peterparker at fastmail.com> [2017-11-19 22:43]:
> > > Hi list,
> > > 
> > > I am trying to understand the coordinates I am getting from pix_blob.
> > > Example patch 03.movement_detection.pd assumes them to be normalized
> > > from 0-1. In my patch the values from pix_blob can easily become
> > > negative or go outside of the 0-1 range. This is especially the case
> > > when the blobs detected should become very large. Has anyone else
> > > discovered this?
> > 
> > This is on Debian testing with Gem 0.93.3 from the Debian package. I
> > tested the same patch on OS X and there pix_blob does output in the 0-1
> > range reliably.
> 
> 
> are the coordinates or the blobsize exceeding the 0-1 for you?
Yes, they do whenever the blob is getting very big or when the detection
seems confused due to a large area of the image being "active".
> For me the y axis is flipped for this patch.
You mean in 03.movement_detection.pd? Yes, I discovered this as well.

cheersio,
P



More information about the Pd-list mailing list