Masking

Masks use grayscale values to control visibility: white areas reveal the layer content, black areas conceal it, and gray values create partial transparency. This system allows editors to hide or show parts of an image without permanently deleting any pixel data. The mask can be modified, refined, or removed entirely at any point, making it the foundation of professional non-destructive editing workflows. Layer masks can also be copied between layers, inverted to create complementary selections, and combined with other masks using blending operations to build complex visibility controls for sophisticated multi-layer compositions.\n\nIn a product photography workflow, a designer might mask a product from its studio background to place it on a white e-commerce backdrop. The mask precisely follows the product's edges, including complex shapes like jewelry chains or fabric folds. If the client later requests a different background color, the designer simply changes the background layer without re-cutting the product.\n\nMask quality directly determines the quality of the final result. A poorly created mask leaves visible halos, rough edges, or missing details like fine hair strands. Professional editors spend significant time refining mask edges, especially around transparent or semi-transparent elements like glass, smoke, and woven fabrics.\n\nMagic Eraser's Background Eraser generates high-quality masks automatically using AI segmentation. The neural network identifies subject boundaries at the pixel level, handling complex edges like hair, fur, and semi-transparent materials that would take minutes to mask manually. The AI produces clean masks suitable for professional compositing, product listing, and design work. Users can apply, adjust, or remove the mask at any point in their workflow, maintaining full creative flexibility without permanently altering the underlying image data. This non-destructive approach means that edits remain fully reversible, which is essential for iterative professional workflows where client feedback may require revisions.