The ability to detect when an image has been manipulated is not a feature that many apps support. Until Adobe came along. It revealed the feature at its ‘Max Sneaks‘ event a few days ago at the company’s annual max creativity conference.
Anchored on its Adobe Sensei AI platform, the Adobe Face Project was previewed at the sneak event where apps are evaluated before they are introduced to the public.
What Is Project Face?
The apps feature audio editing, animation, & photography tools that Adobe subscribers can use to add appropriate lighting edits to images. Instead of applying shadows & a light source to an image based on a 3D globe, the app uses artificial intelligence (AI) & different images to work out the position of the sun, & to apply the appropriate shadows & light during an edit. What’s more, it also makes use of Adobe stock photos & videos to calculate lighting.
Among the many features revealed in the sneak peak, one that was demonstrated during the event was a feature referred to as the ‘Go Figure’. As demonstrated at the event, the app can automatically track how a person moves & translate the movement into animations.
How Does It Work
As a way of living up to its responsibility, the Adobe Face Project is designed to analyze the pixels of an image. It uses a heat map to show which image has been edited, & the probability of manipulation. Hence, the emphasis of the tool is on detecting facial manipulation by generating a probability of manipulation. By making use of a heat map, the tools indicate an edit path on images that are not visible to the natural human eye.
The tool is not only useful as a facial recognition tool but can potentially function as a part of Adobe’s upcoming Content authenticity program. The authenticity program contains features that alert video & audio viewers to unedited & edited imagery. It can also be used to reverse edits to reveal the original image.
Why It Affects Photography?
To take care of the photographer’s issues, the app allows a person to be a part of a group photo even when the person is not around for the group photo. It does this by using Sensei to automatically blend 2 photos in a way that two people can individually participate in shooting using the same background, such that the result is of 2 people being in similar environments.
Looks too hard to process? Look at it this way: with the All In feature, you can act as the photographer for a group shot & take another shot of yourself outside the group shot. Sensei then allows you to blend the images into a single picture.
The event also saw the unveiling of several AI-assisted animation tools like Project Sweet Talk for automatic lip synchronization, Project Go Figure that turns the movement into skeletal frame animations for a virtual character, & Project Sound Seek for eliminating filler words.
Image Credit: Adobe
- “Tokkingheads” uses AI to animate photos – Startups - 2023-06-05
- Branding made easy with “Simplified” – Startups - 2021-06-08
- Send across video proposals instead of text ones – Startups - 2021-06-04
