Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > comp.sys.mac.system > #55220 > unrolled thread
| Started by | Don Bruder <dakidd@sonic.net> |
|---|---|
| First post | 2014-01-23 10:42 -0800 |
| Last post | 2014-01-23 22:06 -0500 |
| Articles | 20 on this page of 145 — 12 participants |
Back to article view | Back to comp.sys.mac.system
Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-23 10:42 -0800
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-23 14:24 -0500
Re: Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-23 13:00 -0800
Re: Any quick, easy way to compare two images? Tom Stiller <clusterfux.upmon@gmail.com> - 2014-01-23 21:11 +0000
Re: Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-23 16:38 -0800
Re: Any quick, easy way to compare two images? john hevelin <bwotte@panix.com> - 2014-01-24 01:57 +0000
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-23 18:03 -0800
Re: Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-23 21:25 -0800
Re: Any quick, easy way to compare two images? JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-01-23 16:31 -0500
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-23 16:57 -0500
Re: Any quick, easy way to compare two images? JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-01-23 17:02 -0500
Re: Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-23 21:28 -0800
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-23 16:32 -0500
Re: Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-23 16:22 -0800
Re: Any quick, easy way to compare two images? Alan Browne <alan.browne@FreelunchVideotron.ca> - 2014-01-23 19:48 -0500
Re: Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-23 23:54 -0800
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-24 08:09 -0800
Re: Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-24 08:44 -0800
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-23 22:06 -0500
Re: Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-23 23:37 -0800
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-23 20:46 +0000
Re: Any quick, easy way to compare two images? Lewis <g.kreme@gmail.com.dontsendmecopies> - 2014-01-23 23:19 +0000
Re: Any quick, easy way to compare two images? Alan Browne <alan.browne@FreelunchVideotron.ca> - 2014-01-23 18:38 -0500
Re: Any quick, easy way to compare two images? Lewis <g.kreme@gmail.com.dontsendmecopies> - 2014-01-24 00:49 +0000
Re: Any quick, easy way to compare two images? Alan Browne <alan.browne@FreelunchVideotron.ca> - 2014-01-23 19:59 -0500
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-23 15:25 -0800
Re: Any quick, easy way to compare two images? Alan Browne <alan.browne@FreelunchVideotron.ca> - 2014-01-23 18:37 -0500
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-23 16:05 -0800
Re: Any quick, easy way to compare two images? Alan Browne <alan.browne@FreelunchVideotron.ca> - 2014-01-23 19:47 -0500
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-23 17:10 -0800
Re: Any quick, easy way to compare two images? Alan Browne <alan.browne@FreelunchVideotron.ca> - 2014-01-24 17:42 -0500
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-24 16:42 -0800
Re: Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-24 20:08 -0800
Re: Any quick, easy way to compare two images? Alan Browne <alan.browne@FreelunchVideotron.ca> - 2014-01-25 10:54 -0500
Re: Any quick, easy way to compare two images? JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-01-25 15:31 -0500
Re: Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-25 17:31 -0800
Re: Any quick, easy way to compare two images? Michael Vilain <vilain@NOspamcop.net> - 2014-01-25 17:57 -0800
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 03:09 +0000
Re: Any quick, easy way to compare two images? JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-01-25 23:02 -0500
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 05:04 +0000
Re: Any quick, easy way to compare two images? Alan Browne <alan.browne@FreelunchVideotron.ca> - 2014-01-25 10:58 -0500
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-25 10:29 -0800
Re: Any quick, easy way to compare two images? Alan Browne <alan.browne@FreelunchVideotron.ca> - 2014-01-25 13:39 -0500
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-25 13:45 -0500
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-25 13:27 -0800
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-25 16:30 -0500
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-25 13:33 -0800
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-25 16:52 -0500
Re: Any quick, easy way to compare two images? JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-01-25 17:31 -0500
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-25 22:35 +0000
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-25 17:49 -0500
Re: Any quick, easy way to compare two images? Tom Stiller <clusterfux.upmon@gmail.com> - 2014-01-25 22:55 +0000
Re: Any quick, easy way to compare two images? JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-01-25 18:33 -0500
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-25 19:42 -0500
Re: Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-25 17:38 -0800
Re: Any quick, easy way to compare two images? JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-01-25 22:50 -0500
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 02:49 +0000
Re: Any quick, easy way to compare two images? Lewis <g.kreme@gmail.com.dontsendmecopies> - 2014-01-26 14:21 +0000
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 15:17 +0000
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-26 14:04 -0500
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 00:17 +0000
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-25 19:42 -0500
Re: Any quick, easy way to compare two images? dorayme <do_ray_me@bigpond.com> - 2014-01-26 12:23 +1100
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 02:54 +0000
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 02:52 +0000
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-25 22:10 -0500
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 03:32 +0000
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-25 22:55 -0500
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 05:02 +0000
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-26 14:04 -0500
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-25 18:03 -0800
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-25 22:10 -0500
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-25 17:49 -0500
Re: Any quick, easy way to compare two images? Savageduck <savageduck1@{REMOVESPAM}me.com> - 2014-01-25 15:59 -0800
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 00:24 +0000
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-25 19:42 -0500
Re: Any quick, easy way to compare two images? Savageduck <savageduck1@{REMOVESPAM}me.com> - 2014-01-25 17:30 -0800
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 02:57 +0000
Re: Any quick, easy way to compare two images? JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-01-25 22:46 -0500
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 03:49 +0000
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-25 22:55 -0500
Re: Any quick, easy way to compare two images? dorayme <do_ray_me@bigpond.com> - 2014-01-26 12:08 +1100
Re: Any quick, easy way to compare two images? Savageduck <savageduck1@{REMOVESPAM}me.com> - 2014-01-25 17:31 -0800
Re: Any quick, easy way to compare two images? dorayme <do_ray_me@bigpond.com> - 2014-01-26 13:00 +1100
Re: Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-25 20:13 -0800
Re: Any quick, easy way to compare two images? dorayme <do_ray_me@bigpond.com> - 2014-01-26 16:15 +1100
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-25 18:07 -0800
Re: Any quick, easy way to compare two images? Savageduck <savageduck1@{REMOVESPAM}me.com> - 2014-01-25 18:24 -0800
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 02:58 +0000
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-25 22:10 -0500
Re: Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-25 17:36 -0800
Re: Any quick, easy way to compare two images? Savageduck <savageduck1@{REMOVESPAM}me.com> - 2014-01-25 17:51 -0800
Re: Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-26 00:39 -0800
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 15:43 +0000
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-26 08:19 -0800
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 16:31 +0000
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-26 09:54 -0800
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 18:14 +0000
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-26 10:44 -0800
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 19:10 +0000
Re: Any quick, easy way to compare two images? dorayme <do_ray_me@bigpond.com> - 2014-01-27 11:14 +1100
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-26 16:22 -0800
Re: Any quick, easy way to compare two images? dorayme <do_ray_me@bigpond.com> - 2014-01-27 11:44 +1100
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-27 13:06 -0500
Re: Any quick, easy way to compare two images? Lewis <g.kreme@gmail.com.dontsendmecopies> - 2014-01-27 12:48 +0000
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-27 05:49 -0800
Re: Any quick, easy way to compare two images? Alan Browne <alan.browne@FreelunchVideotron.ca> - 2014-01-26 12:05 -0500
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 17:08 +0000
Re: Any quick, easy way to compare two images? Alan Browne <alan.browne@FreelunchVideotron.ca> - 2014-01-26 12:27 -0500
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 17:37 +0000
Re: Any quick, easy way to compare two images? JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-01-26 15:36 -0500
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 21:46 +0000
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-26 14:04 -0500
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-26 12:16 -0800
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 21:47 +0000
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-26 16:21 -0800
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-27 01:57 +0000
Re: Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-26 10:32 -0800
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 18:54 +0000
Re: Any quick, easy way to compare two images? JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-01-26 15:40 -0500
Re: Any quick, easy way to compare two images? Savageduck <savageduck1@{REMOVESPAM}me.com> - 2014-01-26 13:30 -0800
Re: Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-26 15:06 -0800
Re: Any quick, easy way to compare two images? JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-01-26 18:20 -0500
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 23:35 +0000
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-27 13:06 -0500
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-27 19:14 +0000
Re: Any quick, easy way to compare two images? JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-01-27 23:18 -0500
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-28 04:27 +0000
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-29 14:01 -0500
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-29 14:01 -0500
Re: Any quick, easy way to compare two images? john hevelin <bwotte@panix.com> - 2014-01-27 01:19 +0000
Re: Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-26 18:28 -0800
Re: Any quick, easy way to compare two images? john hevelin <bwotte@panix.com> - 2014-01-28 02:26 +0000
Re: Any quick, easy way to compare two images? Don Bruder <dakidd@sonic.net> - 2014-01-28 06:35 -0800
Re: Any quick, easy way to compare two images? dorayme <do_ray_me@bigpond.com> - 2014-01-27 10:55 +1100
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-25 22:10 -0500
Re: Any quick, easy way to compare two images? Doug Anderson <ethelthelogremovethis@gmail.com> - 2014-01-25 21:41 -0800
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-26 14:04 -0500
Re: Any quick, easy way to compare two images? Jolly Roger <jollyroger@pobox.com> - 2014-01-26 03:13 +0000
Re: Any quick, easy way to compare two images? JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-01-25 17:30 -0500
Re: Any quick, easy way to compare two images? john hevelin <bwotte@panix.com> - 2014-01-25 18:51 +0000
Re: Any quick, easy way to compare two images? Lewis <g.kreme@gmail.com.dontsendmecopies> - 2014-01-25 19:06 +0000
Re: Any quick, easy way to compare two images? JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-01-25 15:37 -0500
Re: Any quick, easy way to compare two images? JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-01-25 15:35 -0500
Re: Any quick, easy way to compare two images? nospam <nospam@nospam.invalid> - 2014-01-23 22:06 -0500
Page 1 of 8 [1] 2 3 4 5 6 7 8 Next page →
| From | Don Bruder <dakidd@sonic.net> |
|---|---|
| Date | 2014-01-23 10:42 -0800 |
| Subject | Any quick, easy way to compare two images? |
| Message-ID | <lbrnpr$q5o$1@dont-email.me> |
Long story short, I'm messing with an app to assemble a stream of webcam
images into a crude stop-motion movie.
Since the images may arrive at irregular intervals, I'd like to be able
to do something similar to this:
NSImage *PreviousImage; // Global to the class
-(NSImage *)getNewImage:(NSString *)ImageSourceURL
{
NSImage *NewImage = [self grabImageFrom:ImageSourceURL];
if (MAGIC)
{
// Yes, it's a new image
if (PreviousImage)
[PreviousImage release]; // Dump the old one, if any
PreviousImage = [NewImage copy]; // replace with new one
return NewImage; // and send the new one back to caller
}
else
// Nope, image hasn't changed.
return nil;
}
Where "MAGIC" needs to be a routine that returns a YES/NO to the
question "Is NewImage the same as PreviousImage"?
Is there a reasonably painless cocoa/core foundation way to do this, or
am I stuck going right down onto the metal and doing a pixel-by-pixel
comparison "by hand"?
From what I've managed to find in the various cocoa docs, it *LOOKS* as
though the "compare" type methods I can spot decide on equality by
saying "Is this *POINTER* to an NSImage the same as this *OTHER POINTER*
to an NSImage?", without actually looking at the images themselves -
only the pointers to them. Which makes it look as though, should I
assign two identical images to two different NSImages, they'll compare
"not the same", despite the fact that they're identical. Or am I
mis-reading things?
--
Security provided by Mssrs Smith and/or Wesson. Brought to you by the letter Q
[toc] | [next] | [standalone]
| From | nospam <nospam@nospam.invalid> |
|---|---|
| Date | 2014-01-23 14:24 -0500 |
| Message-ID | <230120141424174937%nospam@nospam.invalid> |
| In reply to | #55220 |
In article <lbrnpr$q5o$1@dont-email.me>, Don Bruder <dakidd@sonic.net>
wrote:
> Long story short, I'm messing with an app to assemble a stream of webcam
> images into a crude stop-motion movie.
>
> Since the images may arrive at irregular intervals, I'd like to be able
> to do something similar to this:
>
> NSImage *PreviousImage; // Global to the class
>
> -(NSImage *)getNewImage:(NSString *)ImageSourceURL
> {
> NSImage *NewImage = [self grabImageFrom:ImageSourceURL];
> if (MAGIC)
> {
> // Yes, it's a new image
> if (PreviousImage)
> [PreviousImage release]; // Dump the old one, if any
> PreviousImage = [NewImage copy]; // replace with new one
> return NewImage; // and send the new one back to caller
> }
> else
> // Nope, image hasn't changed.
> return nil;
> }
>
> Where "MAGIC" needs to be a routine that returns a YES/NO to the
> question "Is NewImage the same as PreviousImage"?
>
> Is there a reasonably painless cocoa/core foundation way to do this, or
> am I stuck going right down onto the metal and doing a pixel-by-pixel
> comparison "by hand"?
>
> From what I've managed to find in the various cocoa docs, it *LOOKS* as
> though the "compare" type methods I can spot decide on equality by
> saying "Is this *POINTER* to an NSImage the same as this *OTHER POINTER*
> to an NSImage?", without actually looking at the images themselves -
> only the pointers to them. Which makes it look as though, should I
> assign two identical images to two different NSImages, they'll compare
> "not the same", despite the fact that they're identical. Or am I
> mis-reading things?
the default comparison is by pointer, so you'd need to write your own
equality (and hash) methods. how would it know what you consider
'equal' ? it can't.
also keep in mind that the camera won't produce pixel identical results
on every frame, even when nothing has changed, so you can't just loop
through all pixels and compare. the values will vary slightly.
this is also something that's been solved before, so you probably can
find existing image matching or motion detection code.
[toc] | [prev] | [next] | [standalone]
| From | Don Bruder <dakidd@sonic.net> |
|---|---|
| Date | 2014-01-23 13:00 -0800 |
| Message-ID | <lbrvrh$7ua$1@dont-email.me> |
| In reply to | #55221 |
In article <230120141424174937%nospam@nospam.invalid>,
nospam <nospam@nospam.invalid> wrote:
> In article <lbrnpr$q5o$1@dont-email.me>, Don Bruder <dakidd@sonic.net>
> wrote:
>
> > Long story short, I'm messing with an app to assemble a stream of webcam
> > images into a crude stop-motion movie.
> >
> > Since the images may arrive at irregular intervals, I'd like to be able
> > to do something similar to this:
> >
> > NSImage *PreviousImage; // Global to the class
> >
> > -(NSImage *)getNewImage:(NSString *)ImageSourceURL
> > {
> > NSImage *NewImage = [self grabImageFrom:ImageSourceURL];
> > if (MAGIC)
> > {
> > // Yes, it's a new image
> > if (PreviousImage)
> > [PreviousImage release]; // Dump the old one, if any
> > PreviousImage = [NewImage copy]; // replace with new one
> > return NewImage; // and send the new one back to caller
> > }
> > else
> > // Nope, image hasn't changed.
> > return nil;
> > }
> >
> > Where "MAGIC" needs to be a routine that returns a YES/NO to the
> > question "Is NewImage the same as PreviousImage"?
> >
> > Is there a reasonably painless cocoa/core foundation way to do this, or
> > am I stuck going right down onto the metal and doing a pixel-by-pixel
> > comparison "by hand"?
> >
> > From what I've managed to find in the various cocoa docs, it *LOOKS* as
> > though the "compare" type methods I can spot decide on equality by
> > saying "Is this *POINTER* to an NSImage the same as this *OTHER POINTER*
> > to an NSImage?", without actually looking at the images themselves -
> > only the pointers to them. Which makes it look as though, should I
> > assign two identical images to two different NSImages, they'll compare
> > "not the same", despite the fact that they're identical. Or am I
> > mis-reading things?
>
> the default comparison is by pointer, so you'd need to write your own
> equality (and hash) methods. how would it know what you consider
> 'equal' ? it can't.
That's what I was afraid the docs were saying... <sigh>
> also keep in mind that the camera won't produce pixel identical results
> on every frame, even when nothing has changed,
That alone could be used to tell me "This is a new image". I don't care
if the content of the image has changed, or how, or by how much - I only
care that the two images either are, or are not, two copies of the same
image.
Lemme clarify...
The source I'm using obviously runs a loop something along the lines of
"wake up, get an image from the camera, turn it into a JPG file, and
stick it someplace on the server so the world can get to it, then go to
sleep for a while". The problem is that "for a while" isn't a stable
value - Sometimes it's once per minute, sometimes once per 62 seconds,
sometimes once per second, sometimes once per hour, with no indication
of what condition(s) make it vary. I want to grab every available image,
but NOT grab umpty-nine copies of an image that's been sitting there for
who-knows-how-long
> so you can't just loop
> through all pixels and compare. the values will vary slightly.
I don't see why not - Like I said above, I don't care if the *image
content* changes - only that "the image" I get is or is not the same one
I got the last time I grabbed it.
> this is also something that's been solved before, so you probably can
> find existing image matching or motion detection code.
For what I'm trying to do, image matching and/or motion detection is
likely to be overkill on the order of using a sledgehammer to peel a
hard boiled egg.
If I have to look at the pixels, I expect something as simple as ANDing
them together and calling it "new" if I get back a not-the-same value
from the operation should do the trick.
--
Security provided by Mssrs Smith and/or Wesson. Brought to you by the letter Q
[toc] | [prev] | [next] | [standalone]
| From | Tom Stiller <clusterfux.upmon@gmail.com> |
|---|---|
| Date | 2014-01-23 21:11 +0000 |
| Message-ID | <bkdesoFr87nU1@mid.individual.net> |
| In reply to | #55223 |
On 2014-01-23, Don Bruder <dakidd@sonic.net> wrote: > In article <230120141424174937%nospam@nospam.invalid>, > nospam <nospam@nospam.invalid> wrote: > > If I have to look at the pixels, I expect something as simple as ANDing > them together and calling it "new" if I get back a not-the-same value > from the operation should do the trick. > I'd use XOR and look for a non-zero result to indicate "new".
[toc] | [prev] | [next] | [standalone]
| From | Don Bruder <dakidd@sonic.net> |
|---|---|
| Date | 2014-01-23 16:38 -0800 |
| Message-ID | <lbscm3$7ua$3@dont-email.me> |
| In reply to | #55224 |
In article <bkdesoFr87nU1@mid.individual.net>, Tom Stiller <clusterfux.upmon@gmail.com> wrote: > On 2014-01-23, Don Bruder <dakidd@sonic.net> wrote: > > In article <230120141424174937%nospam@nospam.invalid>, > > nospam <nospam@nospam.invalid> wrote: > > > > If I have to look at the pixels, I expect something as simple as ANDing > > them together and calling it "new" if I get back a not-the-same value > > from the operation should do the trick. > > > I'd use XOR and look for a non-zero result to indicate "new". I think that's likely going to be the way to go - I actually get handed a wad of NSData that I then feed to NSImage to create the actual image, so if I save the raw data to PreviousImage, then XOR it against the "NewImage" data (before feeding it to NSImage), it should be simple enough. Better yet, just grab the same range of bytes from each, rather than doing the full data buffer... Hmmm... Gonna hafta think about this a bit, but I think that's got me pointed in the right direction - I want this to be as "processor lightweight" as I can make it, because I can easily envision wanting to have this program working against multiple webcams, producing multiple movies. -- Security provided by Mssrs Smith and/or Wesson. Brought to you by the letter Q
[toc] | [prev] | [next] | [standalone]
| From | john hevelin <bwotte@panix.com> |
|---|---|
| Date | 2014-01-24 01:57 +0000 |
| Message-ID | <lbshan$kfr$1@reader1.panix.com> |
| In reply to | #55244 |
Don Bruder <dakidd@sonic.net> wrote: > I think that's likely going to be the way to go - I actually get handed > a wad of NSData that I then feed to NSImage to create the actual image, > so if I save the raw data to PreviousImage, then XOR it against the > "NewImage" data (before feeding it to NSImage), it should be simple > enough. Better yet, just grab the same range of bytes from each, rather > than doing the full data buffer... I know nothing about your problem domain, but if what you want to do is compare two sets of bytes for equality, and you don't have metadata such as timestamps, why not use cmp(1)? It does a byte-by-byte comparison and terminates at the first inequality. Your worst case would be if the two sets are identical, since cmp would read both files in their entirety. I wouldn't use diff(1) -- diff would report all the detected differences between the two sets, which could be considerable.
[toc] | [prev] | [next] | [standalone]
| From | Doug Anderson <ethelthelogremovethis@gmail.com> |
|---|---|
| Date | 2014-01-23 18:03 -0800 |
| Message-ID | <x6vbxai1a3.fsf@ethel.the.log> |
| In reply to | #55252 |
john hevelin <bwotte@panix.com> writes: > Don Bruder <dakidd@sonic.net> wrote: > > > I think that's likely going to be the way to go - I actually get handed > > a wad of NSData that I then feed to NSImage to create the actual image, > > so if I save the raw data to PreviousImage, then XOR it against the > > "NewImage" data (before feeding it to NSImage), it should be simple > > enough. Better yet, just grab the same range of bytes from each, rather > > than doing the full data buffer... > > I know nothing about your problem domain, but if what you want to > do is compare two sets of bytes for equality, and you don't have > metadata such as timestamps, why not use cmp(1)? It does a > byte-by-byte comparison and terminates at the first inequality. > Your worst case would be if the two sets are identical, since cmp > would read both files in their entirety. > > I wouldn't use diff(1) -- diff would report all the detected > differences between the two sets, which could be considerable. diff won't do that for binary files (e.g. jpegs), at least not on Os X. It will either report nothing if the files are identical, or will give you a (short) single line of text telling you that they differ if the files are not identical. Of course if they were text files, you could get a lot of output from diff.
[toc] | [prev] | [next] | [standalone]
| From | Don Bruder <dakidd@sonic.net> |
|---|---|
| Date | 2014-01-23 21:25 -0800 |
| Message-ID | <lbstfd$7ua$4@dont-email.me> |
| In reply to | #55224 |
In article <bkdesoFr87nU1@mid.individual.net>, Tom Stiller <clusterfux.upmon@gmail.com> wrote: > On 2014-01-23, Don Bruder <dakidd@sonic.net> wrote: > > In article <230120141424174937%nospam@nospam.invalid>, > > nospam <nospam@nospam.invalid> wrote: > > > > If I have to look at the pixels, I expect something as simple as ANDing > > them together and calling it "new" if I get back a not-the-same value > > from the operation should do the trick. > > > I'd use XOR and look for a non-zero result to indicate "new". Ayup... That's what I was SUPPOSED to say, but something stuttered betwixt brain and fingers :) That's my story, and I'm stickin' to it! :) -- Security provided by Mssrs Smith and/or Wesson. Brought to you by the letter Q
[toc] | [prev] | [next] | [standalone]
| From | JF Mezei <jfmezei.spamnot@vaxination.ca> |
|---|---|
| Date | 2014-01-23 16:31 -0500 |
| Message-ID | <52e18a1b$0$1172$c3e8da3$50776f34@news.astraweb.com> |
| In reply to | #55223 |
On 14-01-23 16:00, Don Bruder wrote: > "wake up, get an image from the camera, turn it into a JPG file, and > stick it someplace on the server so the world can get to it, then go to > sleep for a while". The problem is that "for a while" isn't a stable > value - Sometimes it's once per minute, sometimes once per 62 seconds, Does the .jpg contain any exif data such as date/time of creation ? This would be by far the easiest means to determine it was a new image.
[toc] | [prev] | [next] | [standalone]
| From | nospam <nospam@nospam.invalid> |
|---|---|
| Date | 2014-01-23 16:57 -0500 |
| Message-ID | <230120141657517815%nospam@nospam.invalid> |
| In reply to | #55227 |
In article <52e18a1b$0$1172$c3e8da3$50776f34@news.astraweb.com>, JF Mezei <jfmezei.spamnot@vaxination.ca> wrote: > Does the .jpg contain any exif data such as date/time of creation ? This > would be by far the easiest means to determine it was a new image. not if it's a static scene.
[toc] | [prev] | [next] | [standalone]
| From | JF Mezei <jfmezei.spamnot@vaxination.ca> |
|---|---|
| Date | 2014-01-23 17:02 -0500 |
| Message-ID | <52e19178$0$10612$c3e8da3$33881b6a@news.astraweb.com> |
| In reply to | #55229 |
On 14-01-23 16:57, nospam wrote: > In article <52e18a1b$0$1172$c3e8da3$50776f34@news.astraweb.com>, JF > Mezei <jfmezei.spamnot@vaxination.ca> wrote: > >> Does the .jpg contain any exif data such as date/time of creation ? This >> would be by far the easiest means to determine it was a new image. > > not if it's a static scene. If the camera encodes current time in the EXIF data of generated jpg image, then even if the image is more or less the same, the EXIF data would show different time stamp.
[toc] | [prev] | [next] | [standalone]
| From | Don Bruder <dakidd@sonic.net> |
|---|---|
| Date | 2014-01-23 21:28 -0800 |
| Message-ID | <lbstkb$7ua$5@dont-email.me> |
| In reply to | #55227 |
In article <52e18a1b$0$1172$c3e8da3$50776f34@news.astraweb.com>, JF Mezei <jfmezei.spamnot@vaxination.ca> wrote: > On 14-01-23 16:00, Don Bruder wrote: > > > "wake up, get an image from the camera, turn it into a JPG file, and > > stick it someplace on the server so the world can get to it, then go to > > sleep for a while". The problem is that "for a while" isn't a stable > > value - Sometimes it's once per minute, sometimes once per 62 seconds, > > > Does the .jpg contain any exif data such as date/time of creation ? This > would be by far the easiest means to determine it was a new image. Nope. Not so much as a comment block in it - pure image data. I'd already looked there, but thanks. -- Security provided by Mssrs Smith and/or Wesson. Brought to you by the letter Q
[toc] | [prev] | [next] | [standalone]
| From | nospam <nospam@nospam.invalid> |
|---|---|
| Date | 2014-01-23 16:32 -0500 |
| Message-ID | <230120141632457449%nospam@nospam.invalid> |
| In reply to | #55223 |
In article <lbrvrh$7ua$1@dont-email.me>, Don Bruder <dakidd@sonic.net>
wrote:
> > also keep in mind that the camera won't produce pixel identical results
> > on every frame, even when nothing has changed,
>
> That alone could be used to tell me "This is a new image". I don't care
> if the content of the image has changed, or how, or by how much - I only
> care that the two images either are, or are not, two copies of the same
> image.
>
> Lemme clarify...
> The source I'm using obviously runs a loop something along the lines of
> "wake up, get an image from the camera, turn it into a JPG file, and
> stick it someplace on the server so the world can get to it, then go to
> sleep for a while". The problem is that "for a while" isn't a stable
> value - Sometimes it's once per minute, sometimes once per 62 seconds,
> sometimes once per second, sometimes once per hour, with no indication
> of what condition(s) make it vary. I want to grab every available image,
> but NOT grab umpty-nine copies of an image that's been sitting there for
> who-knows-how-long
>
> > so you can't just loop
> > through all pixels and compare. the values will vary slightly.
>
> I don't see why not - Like I said above, I don't care if the *image
> content* changes - only that "the image" I get is or is not the same one
> I got the last time I grabbed it.
light is random and cameras have a small amount of noise, so each frame
will be slightly different than the previous one, even if nothing has
changed in the scene.
one frame might have a pixel that's {50,100,150} and the next time it's
{51,98,151} and the time after that it's {48,101,153}. that will fail
an equality test even though the images are identical. the brighter the
scene the more it can vary.
completely cover the lens so that no light hits the sensor at all and
the pixels will change slightly from frame to frame (and none will be
{0,0,0} either). there is always going to be some noise.
the difference is below the limits of human perception and you won't
notice it.
[toc] | [prev] | [next] | [standalone]
| From | Don Bruder <dakidd@sonic.net> |
|---|---|
| Date | 2014-01-23 16:22 -0800 |
| Message-ID | <lbsbmr$7ua$2@dont-email.me> |
| In reply to | #55228 |
In article <230120141632457449%nospam@nospam.invalid>,
nospam <nospam@nospam.invalid> wrote:
> In article <lbrvrh$7ua$1@dont-email.me>, Don Bruder <dakidd@sonic.net>
> wrote:
>
> > > also keep in mind that the camera won't produce pixel identical results
> > > on every frame, even when nothing has changed,
> >
> > That alone could be used to tell me "This is a new image". I don't care
> > if the content of the image has changed, or how, or by how much - I only
> > care that the two images either are, or are not, two copies of the same
> > image.
> >
> > Lemme clarify...
> > The source I'm using obviously runs a loop something along the lines of
> > "wake up, get an image from the camera, turn it into a JPG file, and
> > stick it someplace on the server so the world can get to it, then go to
> > sleep for a while". The problem is that "for a while" isn't a stable
> > value - Sometimes it's once per minute, sometimes once per 62 seconds,
> > sometimes once per second, sometimes once per hour, with no indication
> > of what condition(s) make it vary. I want to grab every available image,
> > but NOT grab umpty-nine copies of an image that's been sitting there for
> > who-knows-how-long
> >
> > > so you can't just loop
> > > through all pixels and compare. the values will vary slightly.
> >
> > I don't see why not - Like I said above, I don't care if the *image
> > content* changes - only that "the image" I get is or is not the same one
> > I got the last time I grabbed it.
>
> light is random and cameras have a small amount of noise, so each frame
> will be slightly different than the previous one, even if nothing has
> changed in the scene.
I'm quite thoroughly aware of that fact, thanks. Not only am I aware of
it, *I'M SPECIFICALLY COUNTING ON IT*.
> one frame might have a pixel that's {50,100,150} and the next time it's
> {51,98,151}
Which tells me that the {50,100,150} image has been replaced by the
{51,98,151} image, which in turn tells me that the second one is a new
image. Knowing that, I now know exactly what I'm tying to find out -
Namely, that the second image is not the same image I loaded 5, 10, 50,
500, or 5000 seconds ago.
The content of the image is irrelevant. It can be different, it can be
the same, the differences, if any, might be visible, they might be
imperceptible to a human eye. I just plain don't care. What I'm
interested in - the ONLY thing I'm interested in - is "Did I get a new
image, or did I get another copy of an image I've already seen?"
--
Security provided by Mssrs Smith and/or Wesson. Brought to you by the letter Q
[toc] | [prev] | [next] | [standalone]
| From | Alan Browne <alan.browne@FreelunchVideotron.ca> |
|---|---|
| Date | 2014-01-23 19:48 -0500 |
| Message-ID | <Lcudnbrj4ob7JXzPnZ2dnUVZ8oOdnZ2d@giganews.com> |
| In reply to | #55242 |
On 2014.01.23, 19:22 , Don Bruder wrote:
> In article <230120141632457449%nospam@nospam.invalid>,
> nospam <nospam@nospam.invalid> wrote:
>> light is random and cameras have a small amount of noise, so each frame
>> will be slightly different than the previous one, even if nothing has
>> changed in the scene.
>
> I'm quite thoroughly aware of that fact, thanks. Not only am I aware of
> it, *I'M SPECIFICALLY COUNTING ON IT*.
Then for a movie file with 10000 frames you will get 10000 frames out.
--
“Everybody has a plan until they get punched in the face.”
- Mike Tyson
[toc] | [prev] | [next] | [standalone]
| From | Don Bruder <dakidd@sonic.net> |
|---|---|
| Date | 2014-01-23 23:54 -0800 |
| Message-ID | <lbt677$7ua$7@dont-email.me> |
| In reply to | #55247 |
In article <Lcudnbrj4ob7JXzPnZ2dnUVZ8oOdnZ2d@giganews.com>, Alan Browne <alan.browne@FreelunchVideotron.ca> wrote: > On 2014.01.23, 19:22 , Don Bruder wrote: > > In article <230120141632457449%nospam@nospam.invalid>, > > nospam <nospam@nospam.invalid> wrote: > > >> light is random and cameras have a small amount of noise, so each frame > >> will be slightly different than the previous one, even if nothing has > >> changed in the scene. > > > > I'm quite thoroughly aware of that fact, thanks. Not only am I aware of > > it, *I'M SPECIFICALLY COUNTING ON IT*. > > Then for a movie file with 10000 frames you will get 10000 frames out. Oh Em Gee I'm surrounded by imbeciles. Or trolls. Trolls. Yeah, that's it. It's GOTTA be trolls. PLEASE somebody tell me it's trolls! It's either that, or I just woke up in Bizzarro-land. Please please please please please Puh-LEEZE! god, let it just be trolls... -- Security provided by Mssrs Smith and/or Wesson. Brought to you by the letter Q
[toc] | [prev] | [next] | [standalone]
| From | Doug Anderson <ethelthelogremovethis@gmail.com> |
|---|---|
| Date | 2014-01-24 08:09 -0800 |
| Message-ID | <oh61p9z7hr.fsf@ethel.the.log> |
| In reply to | #55263 |
Don Bruder <dakidd@sonic.net> writes: > In article <Lcudnbrj4ob7JXzPnZ2dnUVZ8oOdnZ2d@giganews.com>, > Alan Browne <alan.browne@FreelunchVideotron.ca> wrote: > > > On 2014.01.23, 19:22 , Don Bruder wrote: > > > In article <230120141632457449%nospam@nospam.invalid>, > > > nospam <nospam@nospam.invalid> wrote: > > > > >> light is random and cameras have a small amount of noise, so each frame > > >> will be slightly different than the previous one, even if nothing has > > >> changed in the scene. > > > > > > I'm quite thoroughly aware of that fact, thanks. Not only am I aware of > > > it, *I'M SPECIFICALLY COUNTING ON IT*. > > > > Then for a movie file with 10000 frames you will get 10000 frames out. > > Oh > > Em > > Gee > > I'm surrounded by imbeciles. Let me point out - you did ask on USENET, so what do you expect! Also, you got at least three useful answers from three different posters (diff, cmp, XOR). Again, that is not such a bad signal to noise ratio for USENET. > Or trolls. Trolls. Yeah, that's it. It's > GOTTA be trolls. PLEASE somebody tell me it's trolls! It's either that, > or I just woke up in Bizzarro-land. USENET. Similar, but not quite the same, to Bizzaro-land.
[toc] | [prev] | [next] | [standalone]
| From | Don Bruder <dakidd@sonic.net> |
|---|---|
| Date | 2014-01-24 08:44 -0800 |
| Message-ID | <lbu58j$mqe$1@dont-email.me> |
| In reply to | #55270 |
In article <oh61p9z7hr.fsf@ethel.the.log>, Doug Anderson <ethelthelogremovethis@gmail.com> wrote: > Don Bruder <dakidd@sonic.net> writes: > > > In article <Lcudnbrj4ob7JXzPnZ2dnUVZ8oOdnZ2d@giganews.com>, > > Alan Browne <alan.browne@FreelunchVideotron.ca> wrote: > > > > > On 2014.01.23, 19:22 , Don Bruder wrote: > > > > In article <230120141632457449%nospam@nospam.invalid>, > > > > nospam <nospam@nospam.invalid> wrote: > > > > > > >> light is random and cameras have a small amount of noise, so each frame > > > >> will be slightly different than the previous one, even if nothing has > > > >> changed in the scene. > > > > > > > > I'm quite thoroughly aware of that fact, thanks. Not only am I aware of > > > > it, *I'M SPECIFICALLY COUNTING ON IT*. > > > > > > Then for a movie file with 10000 frames you will get 10000 frames out. > > > > Oh > > > > Em > > > > Gee > > > > I'm surrounded by imbeciles. > > Let me point out - you did ask on USENET, so what do you expect! > > Also, you got at least three useful answers from three different > posters (diff, cmp, XOR). Again, that is not such a bad signal to > noise ratio for USENET. > > > Or trolls. Trolls. Yeah, that's it. It's > > GOTTA be trolls. PLEASE somebody tell me it's trolls! It's either that, > > or I just woke up in Bizzarro-land. > > USENET. Similar, but not quite the same, to Bizzaro-land. Yeah, I ought to know that by now :) -- Security provided by Mssrs Smith and/or Wesson. Brought to you by the letter Q
[toc] | [prev] | [next] | [standalone]
| From | nospam <nospam@nospam.invalid> |
|---|---|
| Date | 2014-01-23 22:06 -0500 |
| Message-ID | <230120142206409501%nospam@nospam.invalid> |
| In reply to | #55242 |
In article <lbsbmr$7ua$2@dont-email.me>, Don Bruder <dakidd@sonic.net>
wrote:
> > > Lemme clarify...
> > > The source I'm using obviously runs a loop something along the lines of
> > > "wake up, get an image from the camera, turn it into a JPG file, and
> > > stick it someplace on the server so the world can get to it, then go to
> > > sleep for a while". The problem is that "for a while" isn't a stable
> > > value - Sometimes it's once per minute, sometimes once per 62 seconds,
> > > sometimes once per second, sometimes once per hour, with no indication
> > > of what condition(s) make it vary. I want to grab every available image,
> > > but NOT grab umpty-nine copies of an image that's been sitting there for
> > > who-knows-how-long
> > >
> > > > so you can't just loop
> > > > through all pixels and compare. the values will vary slightly.
> > >
> > > I don't see why not - Like I said above, I don't care if the *image
> > > content* changes - only that "the image" I get is or is not the same one
> > > I got the last time I grabbed it.
> >
> > light is random and cameras have a small amount of noise, so each frame
> > will be slightly different than the previous one, even if nothing has
> > changed in the scene.
>
> I'm quite thoroughly aware of that fact, thanks. Not only am I aware of
> it, *I'M SPECIFICALLY COUNTING ON IT*.
>
> > one frame might have a pixel that's {50,100,150} and the next time it's
> > {51,98,151}
>
> Which tells me that the {50,100,150} image has been replaced by the
> {51,98,151} image, which in turn tells me that the second one is a new
> image. Knowing that, I now know exactly what I'm tying to find out -
> Namely, that the second image is not the same image I loaded 5, 10, 50,
> 500, or 5000 seconds ago.
since you said you want to make a stop-motion movie out of it, the
movie will be much bigger than it needs to be since you'll have
zillions of images, all exactly the same.
if you can't see the difference between images, then you might as well
just take one photo and be done with it.
> The content of the image is irrelevant. It can be different, it can be
> the same, the differences, if any, might be visible, they might be
> imperceptible to a human eye. I just plain don't care. What I'm
> interested in - the ONLY thing I'm interested in - is "Did I get a new
> image, or did I get another copy of an image I've already seen?"
you can't tell if you got a copy of an image you already have seen
since every image it takes will be different from the previous images,
even if nothing changed in the scene.
[toc] | [prev] | [next] | [standalone]
| From | Don Bruder <dakidd@sonic.net> |
|---|---|
| Date | 2014-01-23 23:37 -0800 |
| Message-ID | <lbt567$7ua$6@dont-email.me> |
| In reply to | #55256 |
In article <230120142206409501%nospam@nospam.invalid>,
nospam <nospam@nospam.invalid> wrote:
> In article <lbsbmr$7ua$2@dont-email.me>, Don Bruder <dakidd@sonic.net>
> wrote:
>
> > > > Lemme clarify...
> > > > The source I'm using obviously runs a loop something along the lines of
> > > > "wake up, get an image from the camera, turn it into a JPG file, and
> > > > stick it someplace on the server so the world can get to it, then go to
> > > > sleep for a while". The problem is that "for a while" isn't a stable
> > > > value - Sometimes it's once per minute, sometimes once per 62 seconds,
> > > > sometimes once per second, sometimes once per hour, with no indication
> > > > of what condition(s) make it vary. I want to grab every available
> > > > image,
> > > > but NOT grab umpty-nine copies of an image that's been sitting there
> > > > for
> > > > who-knows-how-long
> > > >
> > > > > so you can't just loop
> > > > > through all pixels and compare. the values will vary slightly.
> > > >
> > > > I don't see why not - Like I said above, I don't care if the *image
> > > > content* changes - only that "the image" I get is or is not the same
> > > > one
> > > > I got the last time I grabbed it.
> > >
> > > light is random and cameras have a small amount of noise, so each frame
> > > will be slightly different than the previous one, even if nothing has
> > > changed in the scene.
> >
> > I'm quite thoroughly aware of that fact, thanks. Not only am I aware of
> > it, *I'M SPECIFICALLY COUNTING ON IT*.
> >
> > > one frame might have a pixel that's {50,100,150} and the next time it's
> > > {51,98,151}
> >
> > Which tells me that the {50,100,150} image has been replaced by the
> > {51,98,151} image, which in turn tells me that the second one is a new
> > image. Knowing that, I now know exactly what I'm tying to find out -
> > Namely, that the second image is not the same image I loaded 5, 10, 50,
> > 500, or 5000 seconds ago.
>
> since you said you want to make a stop-motion movie out of it, the
> movie will be much bigger than it needs to be since you'll have
> zillions of images, all exactly the same.
>
> if you can't see the difference between images, then you might as well
> just take one photo and be done with it.
>
> > The content of the image is irrelevant. It can be different, it can be
> > the same, the differences, if any, might be visible, they might be
> > imperceptible to a human eye. I just plain don't care. What I'm
> > interested in - the ONLY thing I'm interested in - is "Did I get a new
> > image, or did I get another copy of an image I've already seen?"
>
> you can't tell if you got a copy of an image you already have seen
> since every image it takes will be different from the previous images,
> even if nothing changed in the scene.
<facepalm>
--
Security provided by Mssrs Smith and/or Wesson. Brought to you by the letter Q
[toc] | [prev] | [next] | [standalone]
Page 1 of 8 [1] 2 3 4 5 6 7 8 Next page →
Back to top | Article view | comp.sys.mac.system
csiph-web