How person selection works
- Upload or open an image layer in the editor and select that layer.
- Choose Lasso cut-out, then select “Auto select person.” On first use, the browser downloads the segmentation model code and model needed to run detection.
- Review the outline. Use the manual lasso and edge softness controls if the boundary needs correction.
- Choose “Cut selection to new layer.” The result remains an editable image layer in the project.
Basic person selection and manual refinement do not use YCORE AI credits. Model download availability and detection quality depend on your browser, network, and source image.
Frequently asked questions
Does it remove every kind of background?
No. Automatic detection targets people. It may not select other objects or produce a clean boundary for every image.
Is my image sent to an AI provider for this selection?
The person segmentation inference runs in the browser. The model assets are fetched from a public content delivery network; the editor does not submit the image to that CDN for segmentation.
Can I fix an imperfect selection?
Yes. Review the outline and use manual lasso controls before cutting the selection to a new layer.