Remember that song from Zager and Evans – ‘In the Year 2525′? Anybody? No? Well look it up kids. Anyway at one point in the year 5555 they say that your legs got nothin’ to do because some machine is doing that for you.
Guess what – in the year 2020 photographers may have nothing to do ’cause some machine is gonna do your photography for you.
Just kidding. It’s already happening.
Check out Arsenal billed as the ‘smart camera assistant’. For a couple of hundred smacks this baby will turn you into a professional photographer. In front of a waterfall or at the beach and want that soft texture in the water? No problem, Arsenal will set you up for a long exposure. Want to stack multiple images for deep depth of field? You got it, Arsenal takes multiple images at various focus points and melds them together. ISO settings, shutter speed, color balance, filters, you name it and this thing does it. It’s part hardware that you attach to your camera and part software that has been trained on thousands of images so that it can determine the settings you need to produce images just like them. You don’t need to do squat. Well you do need to carry it to the spot because it doesn’t have legs – yet.
Here’s another one. Meet Creatism billed as ‘a deep-learning photographer capable of creating professional work’. Technically this isn’t a machine, it’s editing software paired with images from Google Street View. Check out this photo.
Not bad, eh? In one test the Creatism software sampled around 40,000 panoramic images from Street View and picked out about 31,000 sections that were best composed. Then it applied a couple of filters to adjust the exposure, saturation, etc. Finally it applied a special ‘dramatic mask’ to give the photos that professional punch.
The team then showed a mix of Creatism and human created photos to a panel of ‘experts’. Could they tell the computer generated images from the human ones? Nope. The experts ranked the images from 1 to 4 – from looks like a beginner snap to it to looks like a professional created this image with a well thought out composition and intent. They gave this photo 3.5. In fact a significant number of Creatism photos ranked 3.0, semi-professional, or higher – enough so that the developers felt comfortable dubbing the software an actual photographer. Holy Photoshop, Batman! (Of course this could speak more to how bad the experts were at recognizing good photos than the abilities of the software.)
How about this? Besides taking the photo for you and then editing it for you, now you can get software to select, critique, and categorize the final products. Want to know which of your photos are the best? Just ask EyeEm Vision. You can try it out on the website. Just upload an image and EyeEm Vision will rate it for aesthetics and produce a list of possible category tags. Just think, software that tells you if your image is aesthetically pleasing or not.
The Creatism photo above only scored 63% on the aesthetics scale – kind of lukewarm. This post’s featured image fared even worse, only a 59%. Guess EyeEm isn’t into abstract.This one of a cemetery mausoleum scored a 96%. Woo Hoo!
The closeup of a radiator got me a 79%. The photo of a local building at night only got 57%. The software seemed to focus on the text painted on the building and analyzed it based on the font. The one of water reflections scored 68%. EyeEm seemed to be able to figure out that it was taken outdoors, but it thought maybe it was fire or some sort of metal.
After playing with it for a while I get the idea that EyeEm was tuned toward stock photos – one of their example photos of someone’s lunch scored 97% and a woman’s face with blowing hair scored 100%. The sales pitch is aimed at marketing departments that are sorting images for ad campaigns so this makes sense. I assume it can be tuned differently as needed.
So there you have it. Strap a camera with Arsenal on it to a helmet, put it on your gourd, and go for a hike. It will take photos to document your surroundings as you walk. When you get home, run Creatism on them and let it pick out the ‘good’ ones and edit them. Finally use EyeEm to sort and categorize them so that you can toss them on the web to get maximum likes. Hey, maybe someday soon we can all have personal drones that follow us everywhere, document our lives, and post the images without us even thinking about it.
Photography, which from its inception has existed in a constant state of flux, is changing again – radically. So what are we to think about all this?
Well, I for one welcome our new robot overlords. This is exactly what photography needs. My hope is that the rise of machine photography will fracture the medium into two divergent paths that will bring clarity to the confusion that we have now.
As I point out in my post, The Death of Photography, the medium right now is a confusing mess. Photography and the words that surround it have lost all meaning. No one knows what photography is anymore and they certainly can’t pick significant images out from the billions of copy cats posted every day. We are tired of the deluge of images and don’t posses the skills, time, or inclination to discriminate good from bad, real from fake, art from documentation. We mentally close our eyes and let it all flow over us. It’s all one and when everything is photography then nothing is.
My thinking is that when machines take over the production of the vast majority images that we see everyday, we will finally be able to separate them into one big steaming pile of schlock. Everyone will know that this huge pile over here can be safely ignored. No pressure and no confusion – let the torrent of images flow and pay them no heed.
With so many images cleared away, we will finally be able to distinguish the photographs made by real creative artists. No more thousands upon thousands of nearly identical images of Antelope Canyon. No more dead behind the pixels shots of the Cavell Meadows Trail like the one above. These will now be taken by machines and we can ignore them. Instead we will have a manageable amount of significant work that we can feel good about taking the time to study, interact with, enjoy, and posses. Photography as an art form will be reborn.
Of course knowing the source of the photos we view will be all important. We need to be reasonably knowledgeable about who and how our photos are created. This should not be too hard to do.
I read a BBC article today that a popular Instagram star (Instagram star Daryl Aiden Yow used stock image photos) was caught out claiming that stock images and other people’s work were his own. He had over 100,000 followers and Sony even used ‘his’ work in their advertising campaigns for cameras.
A technology editor for Campaign UK magazine, Emily Tan, is quoted in the article as saying:
“What might be worth considering is to really follow and understand an influencer’s personality and integrity before working with them.
“The danger is engaging one purely for their reach and not their authenticity.”
Another thing to point out is that photographers will absolutely need to concentrate on creativity. They probably already are, but just to be sure we all clear on the subject. Once it becomes obvious that machines can competently take most of the photographs that we see today, it will be incumbent upon artists to move the medium forward based completly on creativity.
Machines can analyze millions of past images and learn to mimic them down to minute details. What they can’t do is make the intuitive leaps of the human mind. They can’t think outside their programming. Their work is dead. They will never show us anything new or unique. Only living breathing humans with the divine spark of creativity can do that.
Comments on "Some Machine is Doing That for You"
Wow, didn’t realize these “photographic aids” were out there! The self-driving car comes to photography, in a sense.
I like that point about the experts not necessarily being able to point out a good photo! It’s one of the negative things (among many positive) about photography clubs – self-appointed experts laying out inflexible rules to determine what should be considered an acceptable photograph…. But now we’ve got an app to do that.
Personal drones, documenting our lives…. reminds me of those slightly obsessive conceptual art projects where somebody photographs themselves every day growing old or whatever. Preferably naked (or it wouldn’t be art).
And besides the digital unreliability of the photograph, apparently the video is not far behind…. I read recently that on the dark web, one can find video porn sites where celebrities’ faces and mannerisms are seamlessly joined to writhing bodies not their own. It’s becoming a strange world where you can’t necessarily trust what you see.
These are just a few of the applications out there and more will be coming on line soon. Google is testing something called a GAN (generative adversarial network) that can create a photograph of a ‘person’ from nothing. One neural net starts making an image and the other judges it for how much it looks like a person. They go back and forth until the refined image looks pretty much exactly like a photo of a person.