🟠 Post-ROI Detection Result Alignment

Function Description

This operator is used to align detection results from cropped images that do not maintain original shape back to the original uncropped image through the defined 2D ROI.

Usage Scenarios

In the image detection preprocessing stage, after cropping is performed for target detection, it is necessary to restore the detection results back to the original uncropped image for subsequent processing or visualization analysis.

Input Output

Input

Pre-ROI image: Original color image before cropping.

Detection result: Results obtained from detection on the cropped image, including bounding boxes, classification information, confidence scores and polygon contours.

2D ROI: Two-dimensional rectangular region defining the image cropping area.

Output

Detection result: Detection results aligned to the original image after coordinate system transformation, containing bounding boxes, classification information, confidence scores and polygon contours.