One of the many advantages of digital photography is that you can see what a photo looks like immediately. What happens to that great scenic when flattened out into two dimensions? And is that butterfly lost against the floral background? With this, you still have a chance to re-take the photo, if needed.

Still, there’s always room for the odd and even serendipitous image…

These images, as with most I show in my pages, reflect more or less the original digital image. I’ll usually crop the shots, and I may adjust the shadows and highlights if the picture is too contrasty. Or if there’s dust on the lens, I’ll use the fill tool in Photoshop to ‘erase’ it. In addition, I almost always sharpen an image after making a compressed version of the picture, and always before making a print.

I’m trying out a new Photoshop plugin set, Reindeer’s Optipix, which is said to have superior edge detection and sharpening. When I opened the edge enhancing filter, I was given options to set noise removing radius, and then edge sharpening radius before picking the amount of emphasis. Curious about the impact of these adjustments, I searched around for more detail. In one forum entry about the plugin, the reviewer mentioned that rather than an unsharp mask, which is a Laplacian of the Gaussian, the Optipix filter takes a Difference of Gaussians to enhance edges.

What a blast from the past this was. I remember studying about the Laplacian of the Gaussian (LoG) and Difference of Gaussians (DoG) when I did a college paper on David Marr’s landmark book on computational models of visual neurophysiology, Vision: A Computational Investigation into the Human Representation and Processing of Visual Information. This book had such a profound impact on me that when I graduated from college, I wanted to work with computer vision (in addition to computational linguistics). Of course, when one only has a BS rather than a PhD, one goes to work with databases and FORTRAN rather than computational models of vision.

Still, I have my copy of the book today, and I think I even have that old paper somewhere. I remember when I wrote it out, on an old typerwriter, I had to hand letter in the formulae, and hand draw in the diagrams. My, haven’t we come far? Of course back then, I still had a fair hand and my writing was legible. Now, if it weren’t for my graphics tablet and pen, I’m not sure I’d remember how to hold a pencil.

To return to Vision, in his book, Marr explored various aspects of vision and with each discussed the physiology behind the act, modeling it mathematically, and then deriving a computational representation.

Consider edge detection. Our eyes have ganglion cells that respond differently to the presence or absence of light. ‘On’ centered cells respond when light is first introduced, and continue to fire as long as the light is maintained. Other ‘off’ centered cells respond only when the light is removed. Both actions are necessary to detect an ‘edge’ , which is really nothing more than a light area next to a dark or darker area–a difference of intensity of light.

(Think about tearing open a bag: one hand moves toward you, the other hand moves away and the action results in an effect–the bag opens, the aroma of potato chips fills the air. If both hands moved in the same direction, you’d starve.)

So how does a computer detect an edge? After all it doesn’t have cells.

Well, according to Marr and others of the time, one way to detect edges is to blur the image, and then subtract this from the original in order to determine the changes of intensity or zero crossings; leaving what Marr termed the zero-crossing segments–edges whose slope denotes the level of contrast of a an edge. Computationally, this is equivalent to taking a transform of the aforementioned Difference of Gaussians , and if you view a graph of this equation, it actually physically approximates how one would think of an edge if one imagined it one dimensionally–two steep hills with a deep divide between them.

For someone who alternated between love of mathematics, and terror of the same, Marr’s book was the first time I’d seen a real, rather than ‘accepted’ correlation between complex mathematics and the physical world. What made it ironic is that I didn’t meet this epiphany while studying physics or computer science. No, I was studying about the human visual system in the process of getting a degree in psychology

David, this one's for you

(A zero-crossing drawing, generated by Photoshop.)

Thirty years ago Marr envisioned a time when computers could see and wrote a book called Vision, published after his death of leukemia in 1980 at the age of 35. Five years later, I sat under a tree on campus making notes, and stopping from time to time to just stare at the bark, the birds against the sky, and the shadows the buildings made–crude hand drawings of which would make their way into my report on his research.

Five years later, in 1990, while I wrestled large databases for Boeing, a small company released a product called Photoshop that would incorporate work of Marr and others.

Fifteen years later, I sit with a thin, titanium computer on my lap, Marr’s book turned upside down on the chair’s arm, while I try a plugin downloaded over the Internet that does what I only dreamed could be possible twenty years ago.


Speaking of Technorati

I was asked why Technorati shows me not having updated in 125 days. Since we all know that if I was that long between updates I’d be dead, I have to wonder: does Technorati think me dead?

Seriously, anyone know where this value comes from? I’ve noticed that my links to posts aren’t showing up, either, but assumed this was some cosmic hee hee — like being weblog two thousand and one on Jeneane Sessum’s Aggregator 2.0.


Onions have layers

Here’s an interesting pattern, see if you can spot it:

As regards to the first Chris Lydon radio broadcast, Dave Rogers makes a comment that Mike Sanders really liked but which Doc Searls counters with more positive feedback. Dave then responded to Doc who responded back to Dave.

Dave writes more, and in the middle somewhere points to Jon Garfunkle’s *New GateKeepers –where Jon points to a lot of these same people (Doc, David, Dave, Jarvis, et al). Mike Sanders responds to Dave and Doc and Chris Lydon, and David Weinberger also responds to Dave (linking Doc and Mike) and Doc links everyone for good measure. So do I come to that, but that’s because I’m asking you all: do you see the pattern?

You might be tempted to say that it reminds you of a cat fight but that’s not it–terms like ‘cat fight’ are reserved for disagreements between women, as a way of poking gentle fun at the little ladies and our silly quarrels. It’s used so that people can then approach the discussion with the proper frame of mind.

No one would ever use ‘cat fight’ to describe a serious disagreement among serious participants.

No, the answer I’m looking for is an onion. This conversation reminds me of an onion, and each person contributing is one layer of that onion. Some of the layers are close to the core, others further out — but they’re all onion regardless of relative position, and just as pungent.

Poly saw the pattern. Perhaps we should use the term ‘bear fight’ from now on. A bull fight comes to mind, but that’s already been taken.

*Note that Jon did link to women in the article, and has always been scrupulous at pointing out inequities in the weblogging world. I pointed to his post primarily because it was referenced in this conversation, and I liked the ‘gatekeeper’ title–so he’s not an onion. He’s sort of a scallion.

Critters Photography

Sweet babies and fireflies

The Missouri summer has moved in, with weather in the 80’s, humid, and rich. I’ve moved my walks to the morning, when it’s still cool. Come July and August, even mornings won’t help and that’s when you take the deep, canyon and river hikes.

I went to the zoo to check out the new Fragile Forest exhibit and the baby penguin. Unlike my last trip in the winter, today the place was quite busy, and all the fountains and falls were turned on–I hadn’t realized what a beautiful zoo the St. Louis zoo is. It’s not big, but it is nicely designed, and wonderfully intimate. I guess that Parent Magazine ran a survey and the St. Louis zoo was named the number one zoo for kids in the country–primarily because the various critters are accessible.

The penguin baby was hidden by adults at the Penguin and Puffin exhibit, but it was nice to watch the antics of the birds and to cool off in the 45 degree temperature controlled environment. Unfortunately, the apes were nowhere to be seen at the Fragile Forest, either. It is still too new for the animals, and they spend a lot of time in their old habitat.

However, other animals were out and about and nicely active; including the prairie dog village, which had several babies of their own. I managed to capture a picture of one sweet faced, tiny baby.

The Babe

I really enjoyed the zoo visit today–even taking time to chat with folks, when normally I’m rather shy around strangers. Color, lots of color, and I’ve been of a mood for color. And some excellent fresh cooked, spiced potato chips that I enjoyed by the lake, watching the flamingos.

Flamingo in June

The colorful birds and the antics of the prairie dog pups cheered me considerably. I was in a bit of a dark mood the last few days, which is one reason I wanted to take a break from the computer today. However, as I wrote in comments recently, …a person who is bright and cheerful all the time is on drugs, so at least we know I’m clean and sober.


After today’s flamingos, more color–the fireflies came out tonight. I wish there was a way I capture them on film, but it wouldn’t work. The magic of fireflies is that they glow quickly and just out of the corner of your eye — blinking out when you turn to look. If we captured them on film, the magic would be lost.


Water Bird