Saltar al contenido principal
POST
/
segment-anything-2
Segment Anything 2
curl --request POST \
  --url https://dream-gateway.livepeer.cloud/segment-anything-2 \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: multipart/form-data' \
  --form image='@example-file' \
  --form model_id= \
  --form 'point_coords=<string>' \
  --form 'point_labels=<string>' \
  --form 'box=<string>' \
  --form 'mask_input=<string>' \
  --form multimask_output=true \
  --form return_logits=true \
  --form normalize_coords=true
{
  "masks": "<string>",
  "scores": "<string>",
  "logits": "<string>"
}

Autorizaciones

​
Authorization
string
header
requerido

Bearer authentication header of the form Bearer <token>, where <token> is your auth token.

Cuerpo

multipart/form-data
​
image
file
requerido

Image to segment.

​
model_id
string
predeterminado:""

Hugging Face model ID used for image generation.

​
point_coords
string

Nx2 array of point prompts to the model, where each point is in (X,Y) in pixels.

​
point_labels
string

Labels for the point prompts, where 1 indicates a foreground point and 0 indicates a background point.

​
box
string

A length 4 array given as a box prompt to the model, in XYXY format.

​
mask_input
string

A low-resolution mask input to the model, typically from a previous prediction iteration, with the form 1xHxW (H=W=256 for SAM).

​
multimask_output
boolean
predeterminado:true

If true, the model will return three masks for ambiguous input prompts, often producing better masks than a single prediction.

​
return_logits
boolean
predeterminado:true

If true, returns un-thresholded mask logits instead of a binary mask.

​
normalize_coords
boolean
predeterminado:true

If true, the point coordinates will be normalized to the range [0,1], with point_coords expected to be with respect to image dimensions.

Respuesta

Successful Response

Response model for object segmentation.

​
masks
string
requerido

The generated masks.

​
scores
string
requerido

The model's confidence scores for each generated mask.

​
logits
string
requerido

The raw, unnormalized predictions (logits) for the masks.

Last modified on March 1, 2026