Darknet::Prediction Struct Reference

A much-simplified version of the old API structure DarknetDetection. More...

#include "darknet.hpp"

Collaboration diagram for Darknet::Prediction:

Public Attributes
int	best_class
	Zero-based class index, or -1 if nothing was found in an image.
cv::Point2f	normalized_point
	The center point of the object. This value is normalized and must be multiplied by the image dimensions.
cv::Size2f	normalized_size
	The dimensions of the object. This value is normalized and must be multiplied by the image dimensions.
Probabilities	prob
	The probability for each object. Only non-zero values are kept.
cv::Rect	rect
	The de-normalized bounding box, where the coordinates have been multiplied by the original image width and height.

Detailed Description

A much-simplified version of the old API structure DarknetDetection.

See also

Predictions

DarknetDetection

Since

2024-07-24

Member Data Documentation

best_class

int Darknet::Prediction::best_class

Zero-based class index, or -1 if nothing was found in an image.

normalized_point

cv::Point2f Darknet::Prediction::normalized_point

The center point of the object. This value is normalized and must be multiplied by the image dimensions.

normalized_size

cv::Size2f Darknet::Prediction::normalized_size

The dimensions of the object. This value is normalized and must be multiplied by the image dimensions.

prob

Probabilities Darknet::Prediction::prob

The probability for each object. Only non-zero values are kept.

rect

cv::Rect Darknet::Prediction::rect

The de-normalized bounding box, where the coordinates have been multiplied by the original image width and height.

---

The documentation for this struct was generated from the following file:

• darknet.hpp