Cut a product out of its background
EN
← Back to all capabilities
CAPABILITY 11 · OUTCOME
Image segmentation
Get a mask for one object in an image, selected by a short text label, click points, or a bounding box.
Try Image segmentation with your agentGOOD FOR
Good tasks for this capability
Select one person in a group photo by clicking them
Generate a machine-readable mask for a downstream pipeline
CAPABILITY CONTRACT
Inputs, outputs, and selection criteria
You provide
An image, plus a label, one or more points, or a bounding box marking what to select.
You receive
A masked image file, or a run-length-encoded mask for further processing.
How to choose
Use the image output for previewing or compositing; use RLE when another program needs to consume the mask directly.
SERVICES
Available services now
ServiceReturn modeCatalog price
segment-imagesegment-imageAsyncConfiguration based
segment-image-rlesegment-image-rleAsyncConfiguration based
USE WITH YOUR AGENT
You do not need to pick a model on the website.
After connecting Actionway, tell your local agent the goal. It reads the same service definitions, input schemas, and prices before asking you to approve execution.