nrtk.impls.score_detections.class_agnostic_pixelwise_iou_scorer
This module provides the ClassAgnosticPixelwiseIoUScorer class, which calculates pixelwise Intersection over Union (IoU) scores in a class-agnostic manner. It is intended for evaluating object detection results based on IoU between predicted and actual bounding boxes, without differentiating between object classes.
- Classes:
ClassAgnosticPixelwiseIoUScorer: Computes pixelwise IoU scores for bounding boxes in a class-agnostic fashion.
- Dependencies:
numpy for numerical operations.
smqtk_image_io for handling bounding boxes.
nrtk.interfaces.score_detections.ScoreDetections for the detection scoring interface.
- Example usage:
scorer = ClassAgnosticPixelwiseIoUScorer() scores = scorer.score(actual_detections, predicted_detections)
Classes
Implementation of ScoreDetection interface that computes the Pixelwise IoU scores in a Class-Agnostic manner. |