Adobe Sensei, artificial intelligence to serve creation

Note : This page has been translated into English from French by a machine translation tool

Artificial intelligence is everywhere. If its real or hypothetical contribution may be debatable in some cases, this does not prevent it from being a real plus when it is dedicated to specific fields such as design or audio and video editing. It can simplify the work of professional users as well as private individuals and save them precious time on complex or tedious tasks where their realization by an operator does not strictly speaking provide a significant gain or even none at all. Adobe software companies have been incorporating these services into their products for several years. Let’s take a closer look at what it is all about…

Adobe Sensei is an integrated module in most of Adobe’s software products, introduced in 2016, and based on artificial intelligence (see our Artificial Intelligence article for more information). The name “Sensei” comes from the Japanese word for teacher or master. Sensei combines artificial intelligence and machine learning technologies. Adobe has collected billions of pieces of data from its analyses and images, and through machine learning, they are now able to detect similarities and patterns to produce AI that is effective in their fields.

In each of our articles about Adobe software products, you’ll find the features that are enabled by Sensei. Nevertheless, here are some of the features that Sensei offers in the various Adobe applications :

• Morph Cut in Adobe Premiere Pro analyzes and evaluates the content of a movie or video clip and is able to use Sensei to determine which effect is most appropriate for that video.
• Face Aware Liquify/Editing in Adobe Photoshop Lightroom allows you to apply changes to a face without the changes being applied to the rest of the image when you don’t want them to be applied.
• Sensei optimizes the search for documents by trying to understand their meaning and context in order to make matches, as can be the case between images for example
• Sensei is also a voice-controlled virtual assistant added to the applications and services of the Adobe Creative Cloud software suite

We can imagine that Sensei will eventually allow to fully control Adobe Photoshop by voice, explain to the user how to edit photos and videos like a professional and propel emerging technologies such as virtual reality painting via Project Dali or the photo editing tool Artistic Eye. Adobe, now using computer vision to manipulate photos, aims to extend its vision services to companies or developers who want to analyze images or give visibility to their products. Sensei’s goal is not to generalize its AI but only to offer a variety of services for creative people. In the future, Sensei will help reduce the learning curve and accelerate the completion of creative work.

In an instant, Sensei is able to provide keywords of what he has managed to identify on an image and extract these different parts of the image into thumbnails. He can then propose images from Adobe Stock that are related to the image you are working on. Each keyword is isolated and can be searched again with greater precision. The user can then use sliders to choose the quantity, density, or colors of the item they are looking for to refine the results and find the image that comes closest to what the user is looking for.

But Sensei can also search your own photo library to find a picture of a specific object or character. If you have several photos of what you are looking for, he will let you refine your search characteristics to find your photo among thousands of others. Sensei can create masks on its own to optimize the integration of a photographic object in a photo of its environment. He can even suggest a font that matches the scene’s environment in order to integrate a title in your visual.

Sensei also offers a tool called Creative Graph that allows to reconstruct all the steps that led to the final rendering in the form of a tree and again to apply individual changes on each of the identified elements of the visual without ever having to reselect them via a manual or semi-automated selection tool. Thanks to this system, it is possible to replace an object by another or even a character by another with a single click and with a surprising quality of integration throughout the visual.

Finally, Sensei also offers an API platform that allows business users to upload their own data in order to use Sensei’s services for their analysis and thus benefit from its expertise.