Detections Stitch

Class: DetectionsStitchBlockV1

Source: inference.core.workflows.core_steps.fusion.detections_stitch.v1.DetectionsStitchBlockV1

This block merges detections that were inferred for multiple sub-parts of the same input image into single detection.

Block may be helpful in the following scenarios: * to apply Slicing Adaptive Inference (SAHI) technique, as a final step of procedure, which involves Image Slicer block and model block at previous stages. * to merge together detections performed by precise, high-resolution model applied as secondary model after coarse detection is performed in the first stage and Dynamic Crop is applied later.

Type identifier

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

Properties

Name	Type	Description	Refs
name	str	Enter a unique identifier for this step..	❌
overlap_filtering_strategy	str	Which strategy to employ when filtering overlapping boxes. None does nothing, NMS discards lower-confidence detections, NMM combines them..	✅
iou_threshold	float	Minimum overlap threshold between boxes. If intersection over union (IoU) is above this ratio, discard or merge the lower confidence box..	✅

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 Detections Stitch in version v1.

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

Input and Output Bindings

The available connections depend on its binding kinds. Check what binding kinds Detections Stitch in version v1 has.

Bindings

• input

  • reference_image (image): Original image that was cropped to produce the predictions..
  • predictions (Union[object_detection_prediction, instance_segmentation_prediction]): Model predictions to be merged into the original image..
  • overlap_filtering_strategy (string): Which strategy to employ when filtering overlapping boxes. None does nothing, NMS discards lower-confidence detections, NMM combines them..
  • iou_threshold (float_zero_to_one): Minimum overlap threshold between boxes. If intersection over union (IoU) is above this ratio, discard or merge the lower confidence box..

• output

  • predictions (Union[object_detection_prediction, instance_segmentation_prediction]): Prediction with detected bounding boxes in form of sv.Detections(...) object if object_detection_prediction or Prediction with detected bounding boxes and segmentation masks in form of sv.Detections(...) object if instance_segmentation_prediction.

Example JSON definition of step Detections Stitch in version v1

{
    "name": "<your_step_name_here>",
    "type": "roboflow_core/detections_stitch@v1",
    "reference_image": "$inputs.image",
    "predictions": "$steps.my_object_detection_model.predictions",
    "overlap_filtering_strategy": "nms",
    "iou_threshold": 0.4
}