Yes, I think the photos in the example link are light/white field photography as well rather than masking. Light field photos have a white background and it’s achieved (mostly) in-camera rather than in post-processing by placing a light source behind the subject and usually setting the exposure so that the background is completely white. Sometimes in post-processing you’ll adjust levels (maybe with a mask, but not always) so that the background is truly white. There is dark field photography as well where the background is completely black and this is usually easier than white field because the flash is in front of the subject, on camera or otherwise, and you just set the aperture and exposure to make sure the background is black. Nevertheless white/light field photography isn’t hard, even in the field, it just sometimes – not always – requires a tiny bit more gear to carry around; you can often do white field exposures though just by using the sky
Yep, this is definitely his approach:
"Liittschwager captures more than 135,000 exposures using light and condenser lenses to create the white background that can be seen in most of the images it features. "
A picture says more than a thousand words so I’ve been trying to find the observations I had seen with the added background blur but I’ve got too long of a list of things recently ID’d. I figured it would be easier and faster to create a mock-up using one of my own pictures to illustrate what I came across. So here’s an example.
This is similar to what the observations looked liked. The plant in question was left in focus, the rest was digitally blurred out, possibly using some kind of phone app. They did a little bit nicer job than my quick mock-up here in Photoshop but it was still obvious that it was digital blur and not in-camera bokeh (which would be impossible to do on something like this):
Here’s the original picture for comparison. Without clarification what to look for, this kind of picture would probably linger forever in “Plantae”:
It was the first time I’ve run across it while ID’ing things on iNat so it’s certainly not a common approach. I thought it was actually quite a good way to clarify the ID target in a jumble of things. It works with just what’s in the picture and doesn’t introduce anything extra like boxes or arrows (which in previous discussions have raised concerns regarding confusing the CV algorithm).
If anyone wants to take this example and do a mock-up of their favorite method of clarifying the target and post it here, feel free to do so. Seeing a couple of options illustrated can only help people looking for ideas how to do this.
What I find often helps is to increase the contrast.
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.