The PriorBox plugin layer generates the prior boxes of designated sizes and aspect ratios across all dimensions (H x W). More...

#include <NvInferPluginUtils.h>

Public Attributes
float *	minSize
	Minimum box size in pixels. Can not be nullptr. More...

float *	maxSize
	Maximum box size in pixels. Can be nullptr. More...

float *	aspectRatios
	Aspect ratios of the boxes. Can be nullptr. More...

int32_t	numMinSize
	Number of elements in minSize. Must be larger than 0. More...

int32_t	numMaxSize
	Number of elements in maxSize. Can be 0 or same as numMinSize. More...

int32_t	numAspectRatios
	Number of elements in aspectRatios. Can be 0. More...

bool	flip

bool	clip
	If true, will clip the prior so that it is within [0,1]. More...

float	variance [4]
	Variance for adjusting the prior boxes. More...

int32_t	imgH
	Image height. If 0, then the H dimension of the data tensor will be used. More...

int32_t	imgW
	Image width. If 0, then the W dimension of the data tensor will be used. More...

float	stepH
	Step in H. If 0, then (float)imgH/h will be used where h is the H dimension of the 1st input tensor. More...

float	stepW
	Step in W. If 0, then (float)imgW/w will be used where w is the W dimension of the 1st input tensor. More...

float	offset
	Offset to the top left corner of each cell. More...

Detailed Description

The PriorBox plugin layer generates the prior boxes of designated sizes and aspect ratios across all dimensions (H x W).

PriorBoxParameters defines a set of parameters for creating the PriorBox plugin layer.

Member Data Documentation

◆ aspectRatios

float* nvinfer1::plugin::PriorBoxParameters::aspectRatios

Aspect ratios of the boxes. Can be nullptr.

◆ clip

bool nvinfer1::plugin::PriorBoxParameters::clip

If true, will clip the prior so that it is within [0,1].

◆ flip

bool nvinfer1::plugin::PriorBoxParameters::flip

If true, will flip each aspect ratio. For example, if there is an aspect ratio "r", the aspect ratio "1.0/r" will be generated as well.

◆ imgH

int32_t nvinfer1::plugin::PriorBoxParameters::imgH

Image height. If 0, then the H dimension of the data tensor will be used.

◆ imgW

int32_t nvinfer1::plugin::PriorBoxParameters::imgW

Image width. If 0, then the W dimension of the data tensor will be used.

◆ maxSize

float* nvinfer1::plugin::PriorBoxParameters::maxSize

Maximum box size in pixels. Can be nullptr.

◆ minSize

float* nvinfer1::plugin::PriorBoxParameters::minSize

Minimum box size in pixels. Can not be nullptr.

◆ numAspectRatios

int32_t nvinfer1::plugin::PriorBoxParameters::numAspectRatios

Number of elements in aspectRatios. Can be 0.

◆ numMaxSize

int32_t nvinfer1::plugin::PriorBoxParameters::numMaxSize

Number of elements in maxSize. Can be 0 or same as numMinSize.

◆ numMinSize

int32_t nvinfer1::plugin::PriorBoxParameters::numMinSize

Number of elements in minSize. Must be larger than 0.

◆ offset

float nvinfer1::plugin::PriorBoxParameters::offset

Offset to the top left corner of each cell.

◆ stepH

float nvinfer1::plugin::PriorBoxParameters::stepH

Step in H. If 0, then (float)imgH/h will be used where h is the H dimension of the 1st input tensor.

◆ stepW

float nvinfer1::plugin::PriorBoxParameters::stepW

Step in W. If 0, then (float)imgW/w will be used where w is the W dimension of the 1st input tensor.

◆ variance

float nvinfer1::plugin::PriorBoxParameters::variance[4]

Variance for adjusting the prior boxes.

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

NvInferPluginUtils.h

Public Attributes