Adobe is adding a new set of AI features to its experimental iOS camera application Project Indigo. The focus of the update is not just to generate the graphics, but also to analyse the image, the light and the colour, as well as to reshoot and edit proposals based on existing images. The functionality is still being tested and is open to only some users.
New Photo Point Assessment and Reshoot Proposal
This update adds two types of features related to the video-testing. The first is a “photographic point assessment”, which gives a similar “professional opinion” in terms of structure, light, colour and image emotions. The other category is “porning and editing suggestions”, which suggests how users can retake the same photo, such as adjusting the view, exposing it, or removing interference from the image.
Marc Levoy, who is in charge of the project, indicates that many of the generated AI fixes rely on text tips, but it is often difficult for users to write exactly what they want. Therefore, the project Indigo experiment function is more button-based, giving as much as possible more stable and predictable results.
Remove Object to Process By Category
In the drawing section, Project Indigo also adds a more detailed object removal tool. Unlike common manual circles, the new function allows users to switch directly to the type of content to be removed, such as background persons, garbage, cables, utility poles, fences, vehicles and other miscellaneous items, and also supports the self-defined description of the target audience.
It was mentioned that this function performed more steadily in practical tests, enabling the removal of people and objects in the background while avoiding, to the extent possible, significant image defects. In addition to the removal of objects, the application can also generate a deep-seated image for the photograph by AI, and simulate a false background.
Continue testing style conversion and text editing
At the same time, Adobe tested the style conversion function in its application, which allows users to process photos into different visual styles, such as watercolor, pen drawings, monochromatic or reverse light. These functions are not the first of their kind, but Adobe is trying to improve the quality of the output with more advanced models.
Project Indigo retains the function of “Description of your editorial needs”, despite Adobe's reservations about pure indicative drawings, so that users can complete changes other than preset tools. It was noted that such open editing, while more flexible, was also more likely to produce destabilizing results.
Currently these functions are placed under the application 's AI Playground tab. At this stage, Adobe uses a Google Gemini-based model to support it, while also retaining space to switch to other models, including its own Firefly. The company has not yet indicated whether these experimental functions will be formally rolled out to a wider audience.