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: Google Vision OCR, VLM as Detector, SIFT, Identify Changes, Dynamic Zone, EasyOCR, Icon Visualization, OpenAI, Circle Visualization, Pixelate Visualization, Twilio SMS/MMS Notification, Line Counter Visualization, Detection Event Log, Camera Calibration, Halo Visualization, Relative Static Crop, QR Code Generator, Seg Preview, Identify Outliers, Detections Stabilizer, Moondream2, Bounding Rectangle, Image Slicer, Keypoint Detection Model, Florence-2 Model, Camera Focus, Byte Tracker, Anthropic Claude, OCR Model, Gaze Detection, Email Notification, Mask Visualization, Anthropic Claude, Contrast Equalization, PTZ Tracking (ONVIF).md), SIFT Comparison, Detections Stitch, Instance Segmentation Model, Byte Tracker, Depth Estimation, Google Gemini, Stitch Images, Path Deviation, Template Matching, YOLO-World Model, Florence-2 Model, Triangle Visualization, Label Visualization, LMM, SIFT Comparison, Detection Offset, Text Display, CSV Formatter, Keypoint Visualization, Detections Combine, JSON Parser, VLM as Classifier, Color Visualization, Image Contours, Path Deviation, Roboflow Dataset Upload, Bounding Box Visualization, Morphological Transformation, Detections Classes Replacement, OpenAI, Single-Label Classification Model, Llama 3.2 Vision, Multi-Label Classification Model, Byte Tracker, Roboflow Dataset Upload, Velocity, Model Comparison Visualization, Cosine Similarity, Time in Zone, Google Gemini, Polygon Visualization, Image Slicer, Dynamic Crop, Image Threshold, Local File Sink, Reference Path Visualization, Trace Visualization, Email Notification, Time in Zone, Webhook Sink, Blur Visualization, OpenAI, Ellipse Visualization, Crop Visualization, Polygon Zone Visualization, OpenAI, Grid Visualization, Overlap Filter, Stability AI Inpainting, Detections Transformation, Classification Label Visualization, SAM 3, VLM as Classifier, Instance Segmentation Model, Roboflow Custom Metadata, Stitch OCR Detections, Model Monitoring Inference Aggregator, Image Blur, Corner Visualization, Detections Merge, Object Detection Model, LMM For Classification, Stability AI Outpainting, Time in Zone, SAM 3, Absolute Static Crop, SAM 3, Detections List Roll-Up, VLM as Detector, Background Subtraction, Segment Anything 2 Model, Keypoint Detection Model, Stability AI Image Generation, Line Counter, Detections Consensus, Camera Focus, Detections Filter, Clip Comparison, Twilio SMS Notification, Dot Visualization, CogVLM, Perspective Correction, Image Convert Grayscale, Background Color Visualization, Slack Notification, Motion Detection, Image Preprocessing, Object Detection Model, Anthropic Claude, Google Gemini
• outputs: Roboflow Dataset Upload, Byte Tracker, Velocity, Mask Visualization, Model Comparison Visualization, Time in Zone, Time in Zone, PTZ Tracking (ONVIF).md), Dynamic Zone, Polygon Visualization, Detections List Roll-Up, Dynamic Crop, Detections Stitch, Icon Visualization, Byte Tracker, Byte Tracker, Circle Visualization, Segment Anything 2 Model, Trace Visualization, Path Deviation, Florence-2 Model, Line Counter, Pixelate Visualization, Triangle Visualization, Label Visualization, Detection Event Log, Detections Consensus, Time in Zone, Detections Filter, Halo Visualization, Detection Offset, Blur Visualization, Ellipse Visualization, Dot Visualization, Crop Visualization, Perspective Correction, Detections Combine, Detections Stabilizer, Distance Measurement, Color Visualization, Bounding Rectangle, Corner Visualization, Overlap Filter, Detections Transformation, Stability AI Inpainting, Background Color Visualization, Path Deviation, Roboflow Dataset Upload, Size Measurement, Florence-2 Model, Bounding Box Visualization, Roboflow Custom Metadata, Line Counter, Camera Focus, Detections Classes Replacement, Model Monitoring Inference Aggregator, Detections Merge

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[instance_segmentation_prediction, keypoint_detection_prediction, object_detection_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
}