fast::TensorToBoundingBoxSet class
#include <FAST/Algorithms/NeuralNetwork/TensorToBoundingBoxSet.hpp>

Convert a tensor to a set of bounding boxes.

Contents

Used in BoundingBoxNetwork to convert a tensor to a set of bounding boxes.

Base classes

class ProcessObject
Abstract base class for all process objects.

Public functions

auto create(BoundingBoxNetworkType type, float threshold, std::vector<std::vector<Vector2f>> anchors) -> std::shared_ptr<TensorToBoundingBoxSet>
Create instance.
void setThreshold(float threshold)
void setAnchors(std::vector<std::vector<Vector2f>> anchors)
void setInputConnection(DataChannel::pointer channel) override
void setInputConnection(uint portID, DataChannel::pointer channel) override
void setNrOfInputNodes(int nr)
void loadAttributes() override

Private functions

void execute() override

Function documentation

std::shared_ptr<TensorToBoundingBoxSet> fast::TensorToBoundingBoxSet::create(BoundingBoxNetworkType type, float threshold, std::vector<std::vector<Vector2f>> anchors)

Create instance.

Parameters
type Type of bounding box detection network. Used to determine how the output tensor should be handled to create bounding boxes. Default: YOLOv3
threshold Threshold for how high score a bounding box need to be accepted.
anchors List of anchors to use when calculating bounding boxes parameters from the output tensor.
Returns instance