Adobe has heavily invested in AI, as can be surmised from three new features that are being introduced to its different applications.
The AI revolution is upon us, and much like everything else Hollywood related, the technology is not quite as murder oriented as previously believed. It's just a matter of the hands utilizing said technology. Essentially, AI tech has existed for quite a long time, and machine learning is nothing new. However, it's widespread practical application is a concept that brands and companies have only recently started commercializing and getting comfortable with. After all, even with nuanced machine learning, AI can make so many mistakes and often requires a large amount of input to realize whether or not what it's doing is correct or incorrect. Adobe seems to be having quite a bit of fun with the technology, leading to the features we'll be discussing today.
The company recently held its annual MAX conference, wherein new developments both current and upcoming are revealed and discussed. This year, a series of experimental features by the name of Sneaks were introduced, displaying demos that will be implemented in the future. Some of these Sneaks were uploaded to a Creative Cloud for public access, while others just remained demos limited to the MAX conference. The first of these is simply known as Project Strike a Pose, which is such a cool name to have.
Strike a Pose answers the question: what if you have an image where the subject needs to be positioned differently? This is oftentimes an issue that graphic designers run into: having the perfect model, but not having them in the right pose. Well, Strike a Pose fixes this issue through AI. The AI is fed both the model's photo and any reference of the pose that they need to be positioned in. A target pose, if you will. Then, by mapping out the original model's face and colors onto the target pose, the end result will be achieved: the original model making the pose that you want them to. This is both very useful, but could also lead to some annoying results for the original models in the hands of the wrong people. Which makes the next feature a bit more worrying.
The next feature is labelled Project Morpheus, with the name being a cute allusion to the process of morphing faces and the Morpheus character from The Matrix. Either way, a very appropriate name considering how The Matrix is also about technology leading to disastrous ramifications. While Project Morpheus isn't exactly planning on taking over the world, it does make relatively harmful technology into a commodity: deepfakes.
Morpheus essentially improves upon technology previously introduced to Photoshop and limited to photos. Now, users can take video clips and edit individuals in them to quite the extent. Hair color, glasses, expressions, and poses are only some of the metrics that can be edited using Morpheus. Thankfully, entire faces can't be swapped for each other, and even when implemented correctly, the technology looks just off enough to be recognizable as an edit. However, it still grants technology that can be abused very easily to essentially anyone who either has an Adobe subscription or is an online pirate.
The final Sneak is thankfully a bit more fun and much less stressful when long term ramifications are considered. This final Sneak, adorably labelled Project Sunshine, takes pictures of sketches made on the likes of paper or napkins, and then converts them to digital art. The art isn't color filled in, but an outline of everything is exported over from the taken photo. From there on out, users can edit the drawing as they please, utilizing coloring tools, different pencils and whatnot.
Sneaks aren't full-fledged projects on their own. These are potential features that, if met with a warm community reception, will then be integrated into Adobe's different applications, even under different names sometimes. It's just interesting to see how far AI has come, and just how accessible the technology is.
Read next: Beta Versions of Photoshop and Illustrator Come to Web
The AI revolution is upon us, and much like everything else Hollywood related, the technology is not quite as murder oriented as previously believed. It's just a matter of the hands utilizing said technology. Essentially, AI tech has existed for quite a long time, and machine learning is nothing new. However, it's widespread practical application is a concept that brands and companies have only recently started commercializing and getting comfortable with. After all, even with nuanced machine learning, AI can make so many mistakes and often requires a large amount of input to realize whether or not what it's doing is correct or incorrect. Adobe seems to be having quite a bit of fun with the technology, leading to the features we'll be discussing today.
The company recently held its annual MAX conference, wherein new developments both current and upcoming are revealed and discussed. This year, a series of experimental features by the name of Sneaks were introduced, displaying demos that will be implemented in the future. Some of these Sneaks were uploaded to a Creative Cloud for public access, while others just remained demos limited to the MAX conference. The first of these is simply known as Project Strike a Pose, which is such a cool name to have.
Strike a Pose answers the question: what if you have an image where the subject needs to be positioned differently? This is oftentimes an issue that graphic designers run into: having the perfect model, but not having them in the right pose. Well, Strike a Pose fixes this issue through AI. The AI is fed both the model's photo and any reference of the pose that they need to be positioned in. A target pose, if you will. Then, by mapping out the original model's face and colors onto the target pose, the end result will be achieved: the original model making the pose that you want them to. This is both very useful, but could also lead to some annoying results for the original models in the hands of the wrong people. Which makes the next feature a bit more worrying.
The next feature is labelled Project Morpheus, with the name being a cute allusion to the process of morphing faces and the Morpheus character from The Matrix. Either way, a very appropriate name considering how The Matrix is also about technology leading to disastrous ramifications. While Project Morpheus isn't exactly planning on taking over the world, it does make relatively harmful technology into a commodity: deepfakes.
Morpheus essentially improves upon technology previously introduced to Photoshop and limited to photos. Now, users can take video clips and edit individuals in them to quite the extent. Hair color, glasses, expressions, and poses are only some of the metrics that can be edited using Morpheus. Thankfully, entire faces can't be swapped for each other, and even when implemented correctly, the technology looks just off enough to be recognizable as an edit. However, it still grants technology that can be abused very easily to essentially anyone who either has an Adobe subscription or is an online pirate.
The final Sneak is thankfully a bit more fun and much less stressful when long term ramifications are considered. This final Sneak, adorably labelled Project Sunshine, takes pictures of sketches made on the likes of paper or napkins, and then converts them to digital art. The art isn't color filled in, but an outline of everything is exported over from the taken photo. From there on out, users can edit the drawing as they please, utilizing coloring tools, different pencils and whatnot.
Sneaks aren't full-fledged projects on their own. These are potential features that, if met with a warm community reception, will then be integrated into Adobe's different applications, even under different names sometimes. It's just interesting to see how far AI has come, and just how accessible the technology is.
Read next: Beta Versions of Photoshop and Illustrator Come to Web