Erase Person

The tool removes a person from a photo by applying content-aware inpainting over the selected region. It preserves surrounding textures and lighting to create a plausible final image. The workflow accepts either a mask or a bounding box to specify the target, then generates a seamless fill that integrates with the original scene. Example: remove a person from a crowd shot while keeping shadows and reflections consistent.
The underlying algorithms analyze neighboring pixels, perform patch-based synthesis, and apply blending to minimize seams and color shifts. Optional refinements allow manual brush adjustments to tighten edges or correct artifacts, and an option to preserve key background features such as sky or architectural lines. Advanced detection can handle fine details like hair strands and semi-transparent elements, with configurable radius for blending.
Audience includes professional photographers, retouchers, privacy officers, and social media teams who need to remove bystanders or unwanted people while maintaining visual realism. Typical use cases include redacting sensitive individuals in event photos, preparing stock imagery for licensing, or creating a clean background for composites. The tool is designed for batch-ready outputs and supports common image formats, with quality controls to balance speed and fidelity. Differentiators include a robust inpainting engine trained for urban textures, skin-tone aware blending to avoid artifacts, and a simplified mask workflow that reduces manual effort. Users can start with an automatic erase and then refine with precise masking as needed. Outputs retain metadata where possible and can be exported in standard formats for downstream editing or publishing. Considerations include color space handling (sRGB preferred) and safe comparisons across varied lighting to avoid halo effects.