It doesn’t feel like almost a year ago since Google announced Google Lens. It’s Google’s machine learning system to assist your camera to help make your life easier. It helps it to identify what it’s looking at and then do or show you things based on what it sees. Now, Google has announced that it can see your pet. And not only see them but identify them. The goal is to offer you photo books and videos dedicated to your furry friend.
Portrait Mode has been simultaneously one of the biggest jokes and coolest advancements in smartphone camera technology. Google’s version of it can be found in the portrait mode of the Pixel 2 and Pixel 2 XL smartphones. And they have just released their latest version of it as Open Source, available to any developer who can make use of it.
It’s detailed in Semantic Image Segmentation with DeepLap in Tensorflow on the Google Research blog. And reading how it works is quite interesting, even if you have no idea how to actually do it. Semantic Image Segmentation is basically the process by which pixels in an image are defined by labels, such as “road”, “sky”, “person” or “dog”. It allows apps to figure out what to keep sharp and what to blur.
We’ve all been hearing about AI tech that wants to tell us how good our shots are. Apple recently bought out Regaind to help critique our images. Adobe included something simialr in the newest version of Photoshop Elements 2018. Software solutions so far simply look at the thousands of shots you’ve already made. Google wants to cut out the middleman and put this functionality into the camera.
Google’s new “Clips” camera has quite a few bits missing. It has no LCD, and only one button, a shutter button. Although this button is entirely optional. You see, the camera has AI built into it that uses machine learning to recognise and learn faces. It then seeks out interesting moments to capture all by itself. A little creepy, but also pretty cool.
TechCrunch report that Apple has acquired a small computer vision AI tech startup, Regaind. They say that the report has come from multiple sources, and if true could be very cool for the future of Apple’s mobile photography.
Apple already added an intelligent search to the iPhone Photos app a couple of years ago. It allows you to search for particular things like “tree” or “water”, and will usually give you what you ask for. This acquisition allows Apple to take things to a whole new level, though.
Drone technology has come on so quickly in such a short space of time. Especially the camera technology. I’m not just talking about the quality of the optics and sensors, either. The “brains” behind the visual systems in drones now is just nuts. Even modest consumer drones have facial recognition, subject tracking, and similar features. All these features and help us to achieve the best shot possible.
A team of research from MIT and ETH Zurich have now taken things way beyond that which is currently available to the masses. Building on the basic visual systems, theirs actually allows you to determine where in the frame the subject is positioned. It also lets you choose the camera angle. If you want full frontal, you got it. 3/4 left or right? No problem, it’s just the flick of a menu item.
DeepDream is a computer vision AI created by Google which utilises a convolution neural network. It looks for and enhances patterns in images using a process called algorithmic pareidolia. Essentlly, it’s seeing things that aren’t really there. Like the face we may see on the surface of Mars or bunny rabbits & dragons in clouds.
We’ve seen it used on still images for a while and you can make your own here. But this video takes things to a whole new level. Based on a 5 minute clip from Bob Ross’ The Joy of Painting the visuals in this are just plain ridiculous. And if it wasn’t creepy enough already, the sequence is played backwards. So, have a watch of Bob Ross unpainting a picture on LSD.
Ok, imaging AI is just getting ridiculous now. By now we all know of the Prisma app. It lets you take photos and turn them into images that resemble paintings by artists such as Van Gogh, Monet, etc. A team of reseachers at UC Berkley have come up with a system that does the exact opposite of that. It looks at paintings and turns them into something that resembles a photograph.
Using the same principles, though, the team have also taken things in an entirely different direction, allowing you to map individual objects between each other. Such as turning horses into zebras, or vice versa. And now, we really can compare apples to oranges.
There’s been a lot of buzz lately around AI, “deep learning”, computer vision. It’s all to do with image recognition. Apple and Google have also been implementing it with their mobile operating systems to help categorise your shots. Facebook also does this, too, although it rarely makes itself obvious. It’ll often see faces, and ask if you want to tag somebody. Sometimes it’ll even recognise the person. But that’s about it.
Facebook actually looks at a whole lot more, though. If you want to go digging through page source, you can find this information out yourself. But, it’s a bit of a hassle. Now, a new Chrome extension overlays the information right on top of the image in your browser. While primarily intended for those who use screen readers, it does offer insight into how Facebook automatically reports or censors certain images.
Question: can AI vision systems from Microsoft and Google, which are available for free to anybody, identify NSFW (not safe for work, nudity) images? Can this identification be used to automatically censor images by blacking out or blurring NSFW areas of the image?