Segment Anything 2 Model

Class: SegmentAnything2BlockV1

Source: inference.core.workflows.core_steps.models.foundation.segment_anything2.v1.SegmentAnything2BlockV1

Run Segment Anything 2, a zero-shot instance segmentation model, on an image.

** Dedicated inference server required (GPU recomended) **

You can use pass in boxes/predictions from other models to Segment Anything 2 to use as prompts for the model. If you pass in box detections from another model, the class names of the boxes will be forwarded to the predicted masks. If using the model unprompted, the model will assign integers as class names / ids.

Type identifier

Use the following identifier in step "type" field: roboflow_core/segment_anything@v1to add the block as as step in your workflow.

Properties

Name	Type	Description	Refs
name	str	Enter a unique identifier for this step..	❌
version	str	Model to be used. One of hiera_large, hiera_small, hiera_tiny, hiera_b_plus.	✅
threshold	float	Threshold for predicted masks scores.	✅
multimask_output	bool	Flag to determine whether to use sam2 internal multimask or single mask mode. For ambiguous prompts setting to True is recomended..	✅

The Refs column marks possibility to parametrise the property with dynamic values available in workflow runtime. See Bindings for more info.

Available Connections

Compatible Blocks

Check what blocks you can connect to Segment Anything 2 Model in version v1.

• inputs: Crop Visualization, Grid Visualization, Stability AI Inpainting, Object Detection Model, Detections Merge, LMM For Classification, Google Gemini, Detections Classes Replacement, Byte Tracker, Image Blur, YOLO-World Model, OpenAI, Detections Stabilizer, VLM as Classifier, SIFT, Florence-2 Model, Google Vision OCR, Detections Transformation, Velocity, Keypoint Detection Model, Single-Label Classification Model, Roboflow Dataset Upload, Line Counter, Detections Consensus, Stability AI Outpainting, Perspective Correction, Line Counter Visualization, Dot Visualization, Dynamic Crop, Local File Sink, Llama 3.2 Vision, Instance Segmentation Model, Bounding Box Visualization, Depth Estimation, Image Contours, Detections Filter, Background Color Visualization, Ellipse Visualization, Anthropic Claude, Camera Focus, Roboflow Custom Metadata, Color Visualization, Keypoint Detection Model, Stitch Images, Bounding Rectangle, Moondream2, Roboflow Dataset Upload, Label Visualization, Identify Changes, Relative Static Crop, Byte Tracker, Identify Outliers, SIFT Comparison, Object Detection Model, Reference Path Visualization, Multi-Label Classification Model, Stitch OCR Detections, VLM as Classifier, Webhook Sink, Instance Segmentation Model, Keypoint Visualization, Path Deviation, Gaze Detection, Model Monitoring Inference Aggregator, Slack Notification, Time in Zone, Segment Anything 2 Model, OpenAI, Email Notification, Blur Visualization, Image Threshold, Image Slicer, SIFT Comparison, Dynamic Zone, Image Preprocessing, Stability AI Image Generation, Trace Visualization, Classification Label Visualization, CogVLM, Mask Visualization, Image Convert Grayscale, Path Deviation, PTZ Tracking (ONVIF).md), Overlap Filter, Twilio SMS Notification, Byte Tracker, JSON Parser, Polygon Zone Visualization, Template Matching, OCR Model, CSV Formatter, Camera Calibration, Detections Stitch, Icon Visualization, Model Comparison Visualization, Pixelate Visualization, QR Code Generator, Time in Zone, Florence-2 Model, Cosine Similarity, Image Slicer, Halo Visualization, Absolute Static Crop, VLM as Detector, OpenAI, Detection Offset, Circle Visualization, Corner Visualization, Triangle Visualization, VLM as Detector, LMM, Clip Comparison, Polygon Visualization
• outputs: Byte Tracker, Crop Visualization, Stability AI Inpainting, Detections Merge, Distance Measurement, Byte Tracker, Detections Classes Replacement, Line Counter, Detections Stabilizer, Path Deviation, Florence-2 Model, Model Monitoring Inference Aggregator, Roboflow Dataset Upload, Time in Zone, Segment Anything 2 Model, Detections Transformation, Blur Visualization, Velocity, Dynamic Zone, Trace Visualization, Roboflow Dataset Upload, Line Counter, Path Deviation, Mask Visualization, Overlap Filter, PTZ Tracking (ONVIF).md), Byte Tracker, Detections Consensus, Perspective Correction, Detections Stitch, Icon Visualization, Model Comparison Visualization, Time in Zone, Dot Visualization, Dynamic Crop, Pixelate Visualization, Florence-2 Model, Bounding Box Visualization, Bounding Rectangle, Halo Visualization, Detections Filter, Ellipse Visualization, Detection Offset, Background Color Visualization, Circle Visualization, Size Measurement, Corner Visualization, Triangle Visualization, Roboflow Custom Metadata, Color Visualization, Label Visualization, Polygon Visualization

Input and Output Bindings

The available connections depend on its binding kinds. Check what binding kinds Segment Anything 2 Model in version v1 has.

Bindings

• input

  • images (image): The image to infer on..
  • boxes (Union[keypoint_detection_prediction, object_detection_prediction, instance_segmentation_prediction]): Bounding boxes (from another model) to convert to polygons.
  • version (string): Model to be used. One of hiera_large, hiera_small, hiera_tiny, hiera_b_plus.
  • threshold (float): Threshold for predicted masks scores.
  • multimask_output (boolean): Flag to determine whether to use sam2 internal multimask or single mask mode. For ambiguous prompts setting to True is recomended..

• output

  • predictions (instance_segmentation_prediction): Prediction with detected bounding boxes and segmentation masks in form of sv.Detections(...) object.

Example JSON definition of step Segment Anything 2 Model in version v1

{
    "name": "<your_step_name_here>",
    "type": "roboflow_core/segment_anything@v1",
    "images": "$inputs.image",
    "boxes": "$steps.object_detection_model.predictions",
    "version": "hiera_large",
    "threshold": 0.3,
    "multimask_output": true
}