image-object-picker — Masking AI Model
Object Picker opens the segmentation overlay and lets you click the pieces you actually want, building the mask by hand from what the model found. Single runs only: object indices shift from image to image.
Details
- CategoryMasking
- Year-1
- License
Inputs & Outputs
- ImageInputimage
Upstream segmentation overlay — the colored object mask image where each detected object is painted in a distinct color.
- TextInputstring
Per-object metadata from the upstream segmentation model (class label + (R G B) color + optional bbox per line).
- ImageOutputimage
Binary B&W mask: white where the selected object(s) are, black elsewhere. With mode='invert' the polarity is reversed.
Parameters
- selected_indicesstringdefault
Picked indices (set by picker UI)
Comma-separated zero-based indices of the objects to include (e.g. '0,2,4'), set automatically by the picker UI when the user clicks objects in the modal. The LLM should not set this directly — leave default and let the user pick.
- modestringdefault selectoptions: select, invert
Select / invert
'select' (default) masks only the picked objects. 'invert' masks everything except the picked objects. Use invert for 'remove background' style flows where the picker selects what to keep.
Tags
Alternatives in Image Masking
- Image Masker by Class
Build a B&W mask from segmentation/detection by class name.
- Image Masker by Coord
Build a B&W mask from explicit bbox coordinates.
- Image Masker by Region
Build a B&W mask from line-separated polygons or bounding boxes.
- Manuel Image Masker
Hand-draw a B&W mask with brush tools.
- Mask Grow / Shrink
Grow or shrink a mask by N pixels. Instant.