Class SegmentationParserConfig

Inheritance Relationships

Base Type

Class Documentation

class SegmentationParserConfig : public dai::Buffer

Public Functions

SegmentationParserConfig() = default

Construct SegmentationParserConfig message.

virtual ~SegmentationParserConfig()
void setConfidenceThreshold(float threshold)

Add a confidence threshold to the argmax operation over the segmentation tensor. Pixels with confidence values below this threshold will be assigned the background class (255).

Note

Default is -1.0f, which means no thresholding is applied.

Note

Only applicable if output classes are not in a single layer (eg. classesInOneLayer = false).

Parameters:

threshold – Confidence threshold for segmentation parsing

float getConfidenceThreshold() const

Get confidence threshold

void setStepSize(unsigned int stepSize)

Sets the step size for segmentation parsing. A step size of 1 means every pixel is processed, a step size of 2 means every second pixel is processed, and so on. This can be used to speed up processing at the cost of lower resolution masks.

Parameters:

stepSize – Step size for segmentation parsing

unsigned int getStepSize() const

Gets the step size for segmentation parsing.

virtual void serialize(std::vector<std::uint8_t> &metadata, DatatypeEnum &datatype) const override
inline virtual DatatypeEnum getDatatype() const override

Get the datatype of this specific message.

Returns:

DatatypeEnum

DEPTHAI_SERIALIZE(SegmentationParserConfig, confidenceThreshold, stepSize)

Public Members

float confidenceThreshold = -1.0f
unsigned int stepSize = 1