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

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
}