One of the things I am blogging about today are some photos that have already been posted on the blog. The Photos are slightly different now. When I posted them, they were still photographs. I received a notification on Google Plus that a photo had been posted to my account. The picture showed an animation of the moon phases section of my Shaman Staff Walking Stick changing from waxing crescent to the full moon. The first time a picture was animated I was somewhat puzzled because I had asked for no such thing. Since that first animated photo, I have received some others. The next one had added snow falling to a picture of leafless trees I had posted. Then it added twinkling lights to the Christmas tree in the two photos of the gnome ornaments blog post. (You have to look closely to see them twinkle.)
When I first saw the moon phase animation,I was wondering how someone had accessed my account. Then I clicked on an icon and read about the info about Google Awesome Photos. It turns out, that it was not a someone, it was a something. It was a computer program. It is an intelligent program. And herein lies an opportunity to try to engage Artificial Intelligence (AI). Read on.
The Google photo enhancement program has a feature that can stitch together photos to make panoramas and animate sequences of photos. Yes, many photo enhancement programs can do this. But, this is different in that no one is clicking on instructions telling the program to make a photo brighter, increase the contrast, or that this photo is related to that photo. A computer is doing it on its own. And I will have to say that all in all my photos look better after the computer enhancement. So my question is, just how smart is this program?
If you are an older science fiction reader (as I am), you might remember the book, The Moon is A Harsh Mistress, by Robert Heinlein. In that book, a computer becomes sentient. It starts a conversation with a computer repair person. Back then, AI was all just speculation about the future. Now we have computers that may come to very close to that point. I am just starting to read a futurists book that predicts that we may have AI computers having a significant impact on the global economy as early as 2029, only fifteen years from now. (I've just started reading it. If it turns out to be informative I'll post the title and author. Not all futurist books accurately predict what is happening. Example: Where is my hover car? It was supposed to be in my garage by now.) Anyway, I don't expect this computer to talk to me with real speech at this point, but I am going to try to engage the computer to see if it can extrapolate what I want it to do from the pictures I post. So, this adds a new dimension to a project I had planned for the coming year: the field project.
As regular readers know, I posted a photo last week of the field near my home. I live about a hundred feet from this farm land. I initially had thought to document the changes of this field from winter through the
growing season, autumn, and back to winter as a photography project. I walk nearly every day, and it is always refreshing to see how the field changes from day to day. I had planned to take a picture of the field weekly, but also to take the photo at approximately the same time of day, so that I could also show position of the sun at that time and how it changed over the year. Now I am adding another layer to this project: engaging AI. Can this computer program understand that I want it to make an animation of these photographs to show the sequence of the entire year of the field changing. This is of course doable if you program it to do that, but will it understand by my posting a pictures weekly? Will it recognize the sequence over that period of time? Or does it only analyze the data from a single blog post at once? Or possibly will it only do a few frames of animation and quit? It will be a pretty boring animation if it only uses the first few photos.
So what can I extrapolate from what I've seen from this program so far? The program was able to recognize a sequence of photos as a change occurring in a single object (a moon) in the Shaman Staff posts. It was also able to correctly predict that the lights were twinkling on the Christmas tree. The snow on the barren trees was a close prediction. It posted the photo with the Hatch Tag:Winter. However, in my blog post, I had stated that this was a picture taken in late autumn. So this program was not "reading" what I wrote about it. The program analyzed what it "saw", and determined that this was a winter scene. So, we can understand that the program is smart enough to make assumptions from the data. My digital camera is an older model. At the time of purchase, commercial publicly available GPS was just another futurist idea. Newer cameras have geotags embedded in the data which identifies the location. I would be willing to bet that this computer is smart enough to be able to identify the location even though it does not have geotags from my camera because it could match that picture if other pictures of that scene had been taken and posted from a newer camera. (You should be able to turn off the geotags on your camera. Probably a good idea if you are posting a bunch of personal pictures online. The computer is not the only one seeing that data.) I'm not saying that this is happening to your photos, but that it has the potential to happen. (Think of the defense applications of recognizing location in photos.)
So this is the second in my photo series of this field. It was taken about 8:20 am. I had not really planned on taking the photo so early in the morning when I started the series, but rain was coming in and it was getting ready to pore down. I ran up to the field and snapped some shots and ran back home before the rain started. Today, for the second photo, it was raining when we woke up. I asked my husband, (without any real hope of success) to come out and hold the umbrella for me while I snapped the pictures. It was forty-four degrees Fahrenheit with rain and wind, so I really was not surprised that he quipped, "I feel no need to suffer for your art." So I hoofed it up to the field in the rain to snap some photos for today's blog. It is not a big change from last week at this point. One of these days, I'll get a picture with the sun out. And this is where it ends for today.
Next blog post: Further Projects for 2014.
No comments:
Post a Comment