Skip to main content
joshuaf62375374
Participant
Feb 12, 2026 4:06 PM UTC
Open for Voting

P: Introduce a "Convert Mask to Remove" bridge.

Feature Request: Integration of Masking Selections into Generative Remove

The Problem

Currently, the Generative Remove tool operates in a silo. Even if a user has already generated a precise AI mask (e.g., Select Subject, Select Sky, or Objects) within the Masking panel, they must manually re-brush that same area using the Remove tool’s independent brush. This leads to:

  • Redundant Effort: Painting a mask twice for the same object.

  • Inaccuracy: Manual brushing is often less precise than the AI-powered "Select Subject" logic, leading to poor Generative AI fills.

Proposed Solution

Introduce a "Convert Mask to Remove" bridge. This could be implemented in two ways:

  1. From the Masking Panel: Add a "Remove with Generative AI" option to the "..." menu of any existing mask layer.

  2. Within the Remove Tool: Add a "Use Existing Mask" button that allows users to pick a selection (like Subject or Sky) as the source for the Generative Remove brush.

Use Case Example

A photographer wants to remove a distracting person from a complex background.

  1. They use Select Subject to get a pixel-perfect outline.

  2. Instead of re-painting, they click "Apply Generative Remove to Mask." 3. Lightroom uses the existing high-fidelity mask to inform the Generative AI, resulting in a cleaner edge and a faster workflow.

User Benefit

  • Speed: Eliminates the need for manual brushing on complex subjects.

  • Precision: Leverages the superior edge detection of the Masking engine to ensure the Generative AI knows exactly where the object ends and the background begins.

 

Please feel free to reach out for any clarification. Thank you. 

 

-Josh F.

    2 replies

    HenrikKaalund
    Participating Frequently
    Nov 18, 2026 6:31 PM UTC

    This feature would be very practical and, frankly, obvious. You could, for example, select a person, invert the selection and then remove disturbances within that selection (the background) with the remove tool (Ideally with AI and “detect objects”, if these functions were improved to a point where they would actually remove disturbances instead of creating a lot of new clutter.)

    Translate
    JohanElzenga
    Community Expert
    Community Expert
    Feb 16, 2026 12:53 PM UTC

    While some masking tools may be useful for AI Remove as well, I think you make a fundamental mistake when thinking that you need pixel perfect selections for a perfect AI remove. Generally you don’t want that, because the result will usually be worse, not better. If AI removes a pixel perfect perfect selection, it will inevitably miss a few edge pixels. Shadows and reflections are also usually not included in AI object masks, so these will also not be removed when converting such a mask to an AI Remove selection. But these missed pixels will be used by AI Remove to create a matching new fill, often resulting in a newly created object rather than a fill with background pixels. This is a common user error: when you forget to select the shadow of the object as well, AI Remove often creates a new object to match that shadow.

     

    A good example is ‘people distractions’. If you look at how people are being selected with this function, you’ll notice that the selection is way more loose than what people masks would select. That is not because people distractions uses inferior technology, but because a people mask serves a different purpose than a people distraction selection.
     

    So yes, masking tools might be useful sometimes for AI Remove as well, but converting a mask to a removal selection without taking said differences into account will most likely give disappointing results.

    Translate
    -- Johan W. Elzenga