The Multi-focus approach
Early attempts to improve depth of field in cameras include
reduction, as well as other manipulations of lens aperture,
But with limited success. The FocOz system is based
upon an entirely different principle, namely combination of
several differently focused frames of the same scene.
Early Optical attempt
The history of this technique goes back to (at least) the early 1980's:
A purely optical attempt in this direction (which we saw working) is
described in US Patent 4,741,605, applied for in 1983. Inventors were
well known Swedish entertainer, film director etc. Hans Alfredsson et al:
It was a movie objective which could be focused upon two different
distances from the lens, at the same time. The good news was that a film
maker could now focus upon - say - two different actors. The bad news
was that this optical superposition of images included not only the high
resolution part - but also residual blurred image contributions, which
latter part reduced image contrast intolerably.
Early Electronic efforts
So the question arose about how to get rid of such blurred components.
The early dawn of electronic camera technology gave an answer and
relevant lab experiments were reported in Applied Optics 15 May 1983 p1149
One principle attempted, was to filter away the blurred image 'parts'
by electronic means, exemplified (1984) by United States Patent 4,661,986.
(Burt, Adelson & followers). It meant reduction of blur.
Another method is to
divide these differently focused frames into small image parts, compare
every such little segment with its equivalence in the other frames, select the
best focused ones and reject the (blurred) rest. The selected image parts are
finally reassembled for a mosaic resultant picture. It sounds simple but
it hinges on how to measure 'image sharpness' of each little part correctly.
One straight-away method is to use a well known Auto Focus (AF) principle,
stating that Maximum image contrast equals best focus.
procedure may thus be applied to each individual little picture part, but
disturbances (artefacts) may still occur along edges dividing the scene into
areas of different focus ranges. Methods to detect such edges by means of
Laplace operators and similar filters were described,
of little avail where motifs are lacking such edges. Attempts during the 1990's
were apparently restricted to black and white pictures of relatively small
size. And there wasn't generally enough computer power for the task, in
those days. Struggles against these kinds of problems may well be a reason why
infinite depth of field cameras didn't see the light of day during the 1990's, even
though some progress from microscope and missile applications were reported.
So this was the situation around AD 2000, when we started to develop our
FocOz camera. It wasn't that easy. Regular photography of landscapes involve
objects with very little contrast to measure - like the sky - so we had to invent
new techniques for the FocOz camera. Pixel-contrasts in certain program-
loops are not even measured between individual picture-pixels, but rather against
local intensity averages.
Furthermore all colors are included in one single
algorithm, in effect improving the spatial signal to noise ratio due to increasing
the number of pixels participating in a calculation.
We also considered the fact that frames can be out of focus, relative each other,
to a variable degree. Two frames registered through a FocOz tele-lens are usually
featuring a larger focus difference than in wide angle mode. Here comes
(no retouch!) another FocOz animation where the two focuses are differing so
much that each Mexican figure tends to 'disappear' when out of focus!
It's a hard but successful borderline test for the FocOz camera. Such blur-
differences between dual frames can be quantified and thereafter utilised in
the image processing, contributing considerably to the present FocOz success.
It's not our intention here to extend these glimpses, of the FocOz evolution
into a commercial product, any further. But you may continue reading our two
PCT patents, if you care about further technical details. (Press the 'Inventors
view' button.). We trust you will appreciate the FocOz camera
for at least one good reason: It works as advertised. We convinced ourselves
about that before putting it on the market, as evidenced by the many test-
photos presented . Here comes an other one taken January 20th, 2004:
A somewhat surrealistic super-macro composition, with a strange Winter-
Gardener and his orchid. Actually my wife's live orchid in our livingroom.
But it was 10 degrees below zero (C) and quite wintry in the garden (behind).
The gardener is a 'Kasper theatre' character which I bought at the Christmas
Market in Nuremberg, Germany. The only thing you need for this kind of
strange photography is a FocOz camera and just a little bit of imagination:
Remember: It's a full size 5 Megapixel image, processed with all RGB colours
in one run, with our standard 2,0 GHz PC (Windows XP) computer. It took about
two minutes. Can you wait that long until you have ordered your own FocOz
camera from us?
And what's up next?
It's hard to be a soothsayer - about the future and Depth of Field in
particular. Those FocOz features suggested (below) are not in our
sleeves, but rather in our heads - i.e. described in our patents
(Application Numbers PCT/SE01/02889 and PCT/SE02/01934).
It's not likely that you will encounter such FocOz cameras on the
market in the next few years - just mentioning it, in case you consider
upon sitting down and wait for it.
But we believe this is only the beginning. For example, the FocOz
technology can readily be applied to video cameras.
And it's quite feasible to make cameras where
differently-focused frames are shot simultaneously, so that you may,
indeed, record FocOz pictures of very fast moving objects. It's also
possible to apply the FocOz depth of field technology to the camera
viewfinder itself, so that you may (visually) compose your motif
without any focusing forth and back. In fact no initial focusing
whatsoever will be involved if a set of optimal focuses is
predetermined inside the camera. And the future image processing can
be allocated to the camera itself. It will then be time to bid old
Autofocus goodbye in some standard consumer- cameras: Just press the
button and out comes Infinite Depth of Field photos!