Wow,
that was fast. In my most recent column about the difficulty
of keeping track of your images using cataloguing software,
I wrote:
“... help may be on the way for all of us who find
it hard to keep on tagging photos and that’s a system
being developed, which hopes to perfect software using a
description logic interface
that will scan photos, analyse pixel clusters and detect
various objects in them. Then, if all goes well, the system
will be
able to tag those photos automatically with words that will
describe what they contain.
In the meantime,
however, we’re on our own.”
Apparently,
however, we’re not on our own anymore.
It would appear
that, with the arrival of an interesting and inexpensive
-- $39 US – Windows-only program called photology
there is a way of easily searching your digital images using
a series of easy-to-implement filters.
Now, the system I
was writing about in the previous column was being touted
by another company and I have a feeling
it would be much more elaborate and aimed at professionals,
but
photology seems
quite capable of filling the bill for the point-and-shoot
crowd who only want
to hunt up that photo of the family at the beach two summers
ago.
The interface of photology is simplicity itself. You
don’t
search laboriously file by file or folder by folder, but
instead use filters for such basic things as time of day,
date, year
(which means you better have set your camera properly to
record such basic data). Users can also search by content,
say faces
or flowers or sky or a beach scene.
As well, there are colour
filters that not only give you the basics but allow you to
choose varying shades so that
you can
search for turquoise or teal (without having to know what
those actually are), by selecting from a palette.
As that
if weren’t enough, photology can tell whether
the image is horizontal or vertical and if it was taken inside
or out of doors. It even knows if the picture is in focus
or not.
If you’re into tagging photology will do that
as well. But, really, the whole point of the program, is
so that the
average user can avoid the tagging bloat situation where
you (well, I’m guilty of this too) add tags like “kayak” so
you can distinguish pictures of kayaks from ones tagged with
the word “canoe.”
Tim Lenz, one of the scientist-founders
at Tempe, Arizona’s
Enoetic LCC — which released the first commercial version
of photology in the last week of October — says that
the product is designed to take the anguish (and the procrastination)
out of keeping track of your photos at home.
“We’re really looking for the soccer moms and college
kids and first-time parents and baby boomer grandparents – the
people who take tons and tons of photos and maybe don’t
either have the time or the inclination to do all of the
organizations, but they do want to find their photos,” says
Lenz.
And one thing that the people at photology realize,
says Lenz, is that while a lot of us start out with the best
of
intentions
adding tags to our photos in more elaborate amateur programs
like Google’s free Picassa, our energy soon starts
to lag. Once you hit the thousand picture barrier or so,
it all
becomes too much to wade through.
“Most people start off intending to be organized and
then they get lazy. If you use Picassa you do tagging or
things like
that. But you get this tag cloud explosion. You don’t
remember what tag you used and you get fed up. So you’ve
got a really big SIM card in your camera and you’ve
got a thousand pictures on there, when you download those
you’re
not going to take the time to go through and manually hit
every single one of those.”
While photology is easy for the user to navigate, the actual
underpinnings are far more complex.
“There’s a deceptive amount of stuff going on
under the hood,” says Lenz. “We’ve made
the top level as user-friendly as we think we can because
we believe
the
complexity should be hidden and that’s very different
from most software.”
What’s going on under that
hood, adds Lenz, is a lot of sophisticated numerical processing
algorithms and a lot
of the use of statistics. In order to be able to determine
whether something in a photo is or is not a sunset (or if
its taken inside or outside) the photology team has run through
thousands upon thousands of photos.
“And we drop down, we chop up the photos into bits,
we look at colours, regions and segmentation. We look at
individual
faces.”
While photology can tell if there are people
in the photos it examines it doesn’t go as far as actually
recognizing individuals, like, say, your Aunt Sally.
“That would require users to go through and do manual
training,” says
Lenz. “They’d have to have pictures of Aunt
Sally and tell [the program] that those are Aunt Sally.”
And then, says Lenz, all you’d be able to do with
all that training is identify the remaining three pictures
that
you have that are Aunt Sally.
While the photology team is
made up of scientists who are quite capable of coming up
with a more elaborate program
that’s
not what they decided to do here.
“Our background is high-end scientific instrumentation,
so we’re
very comfortable doing the very complicated things. But that’s
a slippery slope, because once it gets really complicated
then the users feel that they should be able to determine
exactly
what feature set is embodied in the software and that becomes
a wild goose chase.”
One thing that Lenz emphasizes
is that photology is not a replacement for a program like
iPhoto.
“We realize that you can do a lot more with photos
in something like iPhoto than you can with photology and
that’s okay.
Our expertise is not in making calendars and laying out books
and printing stamps and all that kind of stuff. We want to
provide a service that we feel doesn’t exist out there.
There’s nobody in the industry that allows you to search
for your photos based on the content in the photos without
you having to go through and manually identify the content
yourself.”
If you’re interested in taking a look at photology then
you’re in for a treat when you visit the web site,
because it comes with a solid explanation of how the software
works
along with several tutorials that will help you get a complete
idea of just how the software is supposed to work.
Oh, and if you’re a Mac user then there will likely
be an OS X version of photology sometime in the feature,
although
not for several months.
The line between the digital camera and the cell phone became
even more blurred recently, with the announcement by Samsung
that it has come out with its new G800 handset, which it
claims with the world’s first 5.0 megapixel camera
phone with 3x optical zoom.
The G800 has the appearance of an everyday digital camera
from one side, while on the reverse it looks like a modern
mobile
phone.
As well as the still camera, the G800 offers video
capture capabilities and face detection.
It’s not surprising
that the owners of digital SLRs buy more photo equipment,
but what might be a bit of a shock
is
just how much more. And considering the figures are based
on 2006, this will likely only get better.
According to figures from PMA Marketing research, 41 per
cent of DSLR owners reported buying a new camera in 2006
vs. 20
per cent of point and shooters. When it came to photo printers,
23 per cent of the DSLR folks bought one as against 14
per cent of the plain Jane digital camera people.
The figures
continue to be impressive for digital picture frames (13
per cent vs six per cent), online storage (four
per cent
vs. one per cent) and photo books (17 per cent vs. seven
per cent.)
PMA has said that camera stores should, therefore
create more bundles for DLSR owners, who are not in any
particular demographic
grouping but spread across age and gender.
Links to Peter's previous commentaries:
|
is
part of the 