Photographic style transfer & Adobe go way back. In 2003 we shipped Match Color, and in 2006 researcher Sylvain Paris was posting results that we sneak-peeked in 2010 and again later as the tech evolved. Now the company is working with Cornell on a deep-learning-based approach.
PetaPixel provides fun examples & details:
[W]hen you put in two photographs, the neural network-powered program analyzes the color and quality of light in the reference photo, and pastes that photo’s characteristics onto the second. This includes things like weather, season, and time of day—theoretically, a winter’s day can be turned into summer, or a cloudy day into a glorious sunrise. […]
It’s important to note that the software does not alter the structure of the photo in any way, so there’s no risk of distorting the lines, edges or perspective. The entire focus is on mimicking the color and light in order to copy the “look” or “style” of a reference photograph onto a new shot.
It’ll be fun to take this next generation for a spin (hopefully soon!).
On a related (?) note, I enjoyed this tweet: “This is the first thing i’ve seen an AI do that has truly terrified me: [see below]”
Back at Adobe, Aseem Agarwala was one of my favorite collaborators. He helped teams invent & implement everything from Photoshop’s layer alignment/blending & Perspective Warp to Warp Stabilizer in After Effects. For the past two years he’s been at Google, now teaming up with researchers to produce Video Frame Synthesis using Deep Voxel Flow. Here, just take a look:
Heh—dig this short animation “based on everyday scenarios we experience through our daily adventures in Los Angeles,” and check out stills/GIFs on Behance.
[Vimeo] [Via Margot Nack]
Little Inventors “takes children’s invention ideas and makes them real.
Children can freely upload their ideas to this website for positive and encouraging feedback from the Little Inventors team… Our aim is to inspire and support children around the world to use their wonderful imaginations to think up ingenious, fantastical, funny or perfectly practical invention ideas with no limits!
Don’t miss this guy’s sweet snack-hat. 🙂
Elsewhere, in The Monster Project (recently exhibited at Adobe SF), “Kids draw monsters, then artists from all over the world recreate them in their own styles!” Check out their fun gallery.
Rather cool! Now I kind of want to buy a pair of these (one for my distant parents, one for us in CA, so that the kids & my mom can trade drawings—one on each fridge):
Joto draws live on request, sketching everything from works of art to messages.
Using the Joto app, on a computer or mobile device, you can discover, create and share jots. These are some ways you can turn images and text into beautiful ink drawings and make the most of Joto.
This is coming. We’re all going over the cliff together.
“Strange Beasts” is an augmented reality game. It allows you to create and grow your own virtual pet. How far can it go?
Check out this rather brilliant Kickstarter project from some Stanford scientists on a mission to broaden access to the wonders of exploring our world:
Foldscope is a real microscope, with magnification and resolution sufficient for imaging live individual cells, cellular organelles, embryos, swimming bacteria and much more. Because the Foldscope is so affordable and can be used anywhere, it brings science to your daily life, whether that means looking at what’s growing in your flower pot or watching bacteria from your mouth or analysing the bee stinger that got your thumb. Our goal is to encourage and enable the curious explorer in each of us and make science happen anywhere, anytime.
Wanna actually go to Mars & feel perpetually jetlagged? Hmm—while thinking that over, take a beautifully painterly flight over the planet surface, courtesy of Jan Fröjdman working with real NASA data:
The anaglyph images of Mars taken by the HiRISE camera holds information about the topography of Mars surface. There are hundreds of high-resolution images of this type. This gives the opportunity to create different studies in 3D. In this film I have chosen some locations and processed the images into panning video clips…
It has really been time-consuming making these panning clips. In my 3D-process I have manually hand-picked reference points on the anaglyph image pairs… The colors in this film are false because the anaglyph images are based on grayscale images. I have therefore color graded the clips. But I have tried to be moderate doing this.
It’s not glamorous, but optimizing apps for low-bandwidth environments is critical to democratizing access to their benefits. Having traveled in Nepal, I can tell you that all the cool creations in the world don’t matter if you can’t even back up & share your photos.
With that in mind, Google Photos is rolling out some important updates:
Now your photos will back up automatically in a lightweight preview quality that’s fast on 2G connections and still looks great on a smartphone. And when a good Wi-Fi connection becomes available, your backed up photos will be replaced with high-quality versions. We’re also making it easier to share many photos at once even on low connectivity. Never mind if you’re at the beach or hiking in the mountains, with Google Photos you can now share pictures quickly even with a spotty connection by sending first in low resolution so friends and family can view them right away. They’ll later update in higher resolution when connectivity permits.
The team esta en fuego!. On the heels of recent releases that added sharable looks, curves, and more, comes some new hotness:
Snapseed 2.17 starts rolling out today and it brings you three new awesome tools:
- Double Exposure allows you to blend two photos and choose from blending modes that are inspired by analog film techniques as well as digital image processing.
- Face Pose lets you correct the pose of portraits based on three dimensional models.
- Expand allows you to increase the size of your canvas and fill up the new space in smart ways with content from your image.
Enjoy, and as always, please let us know what you think.