Object detection DNN (SSD, DetectNet) More...

Classes
class	detectNet
	Object recognition and localization networks with TensorRT support. More...

Macros
#define	DETECTNET_DEFAULT_INPUT "data"
	Name of default input blob for DetectNet caffe model. More...

#define	DETECTNET_DEFAULT_COVERAGE "coverage"
	Name of default output blob of the coverage map for DetectNet caffe model. More...

#define	DETECTNET_DEFAULT_BBOX "bboxes"
	Name of default output blob of the grid of bounding boxes for DetectNet caffe model. More...

#define	DETECTNET_DEFAULT_CONFIDENCE_THRESHOLD 0.5f
	Default value of the minimum detection threshold. More...

#define	DETECTNET_DEFAULT_CLUSTERING_THRESHOLD 0.75f
	Default value of the clustering area-of-overlap threshold. More...

#define	DETECTNET_DEFAULT_THRESHOLD DETECTNET_DEFAULT_CONFIDENCE_THRESHOLD
	Default value of the minimum detection threshold. More...

#define	DETECTNET_DEFAULT_ALPHA 120
	Default alpha blending value used during overlay. More...

#define	DETECTNET_MODEL_TYPE "detection"
	The model type for detectNet in data/networks/models.json. More...

#define	DETECTNET_USAGE_STRING
	Standard command-line options able to be passed to detectNet::Create() More...

Detailed Description

Object detection DNN (SSD, DetectNet)

Class Documentation

◆ detectNet

class detectNet

Object recognition and localization networks with TensorRT support.

Inheritance diagram for detectNet:

Public Types
enum	OverlayFlags { OVERLAY_NONE = 0, OVERLAY_BOX = (1 << 0), OVERLAY_LABEL = (1 << 1), OVERLAY_CONFIDENCE = (1 << 2), OVERLAY_TRACKING = (1 << 3), OVERLAY_LINES = (1 << 4), OVERLAY_DEFAULT = OVERLAY_BOX\|OVERLAY_LABEL\|OVERLAY_CONFIDENCE }
	Overlay flags (can be OR'd together). More...

Public Member Functions
virtual	~detectNet ()
	Destory. More...

template<typename T >
int	Detect (T image, uint32_t width, uint32_t height, Detection *detections, uint32_t overlay=OVERLAY_DEFAULT)
	Detect object locations from an image, returning an array containing the detection results. More...

template<typename T >
int	Detect (T image, uint32_t width, uint32_t height, Detection detections, uint32_t overlay=OVERLAY_DEFAULT)
	Detect object locations in an image, into an array of the results allocated by the user. More...

int	Detect (void input, uint32_t width, uint32_t height, imageFormat format, Detection *detections, uint32_t overlay=OVERLAY_DEFAULT)
	Detect object locations from an image, returning an array containing the detection results. More...

int	Detect (void input, uint32_t width, uint32_t height, imageFormat format, Detection detections, uint32_t overlay=OVERLAY_DEFAULT)
	Detect object locations from an image, into an array of the results allocated by the user. More...

int	Detect (float input, uint32_t width, uint32_t height, Detection *detections, uint32_t overlay=OVERLAY_DEFAULT)
	Detect object locations from an RGBA image, returning an array containing the detection results. More...

int	Detect (float input, uint32_t width, uint32_t height, Detection detections, uint32_t overlay=OVERLAY_DEFAULT)
	Detect object locations in an RGBA image, into an array of the results allocated by the user. More...

template<typename T >
bool	Overlay (T input, T output, uint32_t width, uint32_t height, Detection *detections, uint32_t numDetections, uint32_t flags=OVERLAY_DEFAULT)
	Draw the detected bounding boxes overlayed on an RGBA image. More...

bool	Overlay (void input, void output, uint32_t width, uint32_t height, imageFormat format, Detection *detections, uint32_t numDetections, uint32_t flags=OVERLAY_DEFAULT)
	Draw the detected bounding boxes overlayed on an RGBA image. More...

float	GetThreshold () const
	Retrieve the minimum threshold for detection. More...

void	SetThreshold (float threshold)
	Set the minimum threshold for detection. More...

float	GetConfidenceThreshold () const
	Retrieve the minimum threshold for detection. More...

void	SetConfidenceThreshold (float threshold)
	Set the minimum threshold for detection. More...

float	GetClusteringThreshold () const
	Retrieve the overlapping area % threshold for clustering. More...

void	SetClusteringThreshold (float threshold)
	Set the overlapping area % threshold for clustering. More...

objectTracker *	GetTracker () const
	Get the object tracker being used. More...

void	SetTracker (objectTracker *tracker)
	Set the object tracker to be used. More...

uint32_t	GetMaxDetections () const
	Retrieve the maximum number of simultaneous detections the network supports. More...

uint32_t	GetNumClasses () const
	Retrieve the number of object classes supported in the detector. More...

const char *	GetClassLabel (uint32_t index) const
	Retrieve the description of a particular class. More...

const char *	GetClassDesc (uint32_t index) const
	Retrieve the description of a particular class. More...

const char *	GetClassSynset (uint32_t index) const
	Retrieve the class synset category of a particular class. More...

const char *	GetClassPath () const
	Retrieve the path to the file containing the class descriptions. More...

float4	GetClassColor (uint32_t classIndex) const
	Retrieve the RGBA visualization color a particular class. More...

void	SetClassColor (uint32_t classIndex, const float4 &color)
	Set the visualization color of a particular class of object. More...

void	SetClassColor (uint32_t classIndex, float r, float g, float b, float a=255.0f)
	Set the visualization color of a particular class of object. More...

float	GetLineWidth () const
	Retrieve the line width used during overlay when OVERLAY_LINES is used. More...

void	SetLineWidth (float width)
	Set the line width used during overlay when OVERLAY_LINES is used. More...

float	GetOverlayAlpha () const
	Retrieve the overlay alpha blending value for classes that don't have it explicitly set (between 0-255). More...

void	SetOverlayAlpha (float alpha)
	Set overlay alpha blending value for all classes (between 0-255). More...

Public Member Functions inherited from tensorNet
virtual	~tensorNet ()
	Destory. More...

bool	LoadNetwork (const char prototxt, const char model, const char mean=NULL, const char input_blob="data", const char output_blob="prob", uint32_t maxBatchSize=DEFAULT_MAX_BATCH_SIZE, precisionType precision=TYPE_FASTEST, deviceType device=DEVICE_GPU, bool allowGPUFallback=true, nvinfer1::IInt8Calibrator calibrator=NULL, cudaStream_t stream=NULL)
	Load a new network instance. More...

bool	LoadNetwork (const char prototxt, const char model, const char mean, const char input_blob, const std::vector< std::string > &output_blobs, uint32_t maxBatchSize=DEFAULT_MAX_BATCH_SIZE, precisionType precision=TYPE_FASTEST, deviceType device=DEVICE_GPU, bool allowGPUFallback=true, nvinfer1::IInt8Calibrator *calibrator=NULL, cudaStream_t stream=NULL)
	Load a new network instance with multiple output layers. More...

bool	LoadNetwork (const char prototxt, const char model, const char mean, const std::vector< std::string > &input_blobs, const std::vector< std::string > &output_blobs, uint32_t maxBatchSize=DEFAULT_MAX_BATCH_SIZE, precisionType precision=TYPE_FASTEST, deviceType device=DEVICE_GPU, bool allowGPUFallback=true, nvinfer1::IInt8Calibrator calibrator=NULL, cudaStream_t stream=NULL)
	Load a new network instance with multiple input layers. More...

bool	LoadNetwork (const char prototxt, const char model, const char mean, const char input_blob, const Dims3 &input_dims, const std::vector< std::string > &output_blobs, uint32_t maxBatchSize=DEFAULT_MAX_BATCH_SIZE, precisionType precision=TYPE_FASTEST, deviceType device=DEVICE_GPU, bool allowGPUFallback=true, nvinfer1::IInt8Calibrator *calibrator=NULL, cudaStream_t stream=NULL)
	Load a new network instance (this variant is used for UFF models) More...

bool	LoadNetwork (const char prototxt, const char model, const char mean, const std::vector< std::string > &input_blobs, const std::vector< Dims3 > &input_dims, const std::vector< std::string > &output_blobs, uint32_t maxBatchSize=DEFAULT_MAX_BATCH_SIZE, precisionType precision=TYPE_FASTEST, deviceType device=DEVICE_GPU, bool allowGPUFallback=true, nvinfer1::IInt8Calibrator calibrator=NULL, cudaStream_t stream=NULL)
	Load a new network instance with multiple input layers (used for UFF models) More...

bool	LoadEngine (const char engine_filename, const std::vector< std::string > &input_blobs, const std::vector< std::string > &output_blobs, nvinfer1::IPluginFactory pluginFactory=NULL, deviceType device=DEVICE_GPU, cudaStream_t stream=NULL)
	Load a network instance from a serialized engine plan file. More...

bool	LoadEngine (char engine_stream, size_t engine_size, const std::vector< std::string > &input_blobs, const std::vector< std::string > &output_blobs, nvinfer1::IPluginFactory pluginFactory=NULL, deviceType device=DEVICE_GPU, cudaStream_t stream=NULL)
	Load a network instance from a serialized engine plan file. More...

bool	LoadEngine (nvinfer1::ICudaEngine *engine, const std::vector< std::string > &input_blobs, const std::vector< std::string > &output_blobs, deviceType device=DEVICE_GPU, cudaStream_t stream=NULL)
	Load network resources from an existing TensorRT engine instance. More...

bool	LoadEngine (const char filename, char stream, size_t size)
	Load a serialized engine plan file into memory. More...

void	EnableLayerProfiler ()
	Manually enable layer profiling times. More...

void	EnableDebug ()
	Manually enable debug messages and synchronization. More...

bool	AllowGPUFallback () const
	Return true if GPU fallback is enabled. More...

deviceType	GetDevice () const
	Retrieve the device being used for execution. More...

precisionType	GetPrecision () const
	Retrieve the type of precision being used. More...

bool	IsPrecision (precisionType type) const
	Check if a particular precision is being used. More...

cudaStream_t	GetStream () const
	Retrieve the stream that the device is operating on. More...

cudaStream_t	CreateStream (bool nonBlocking=true)
	Create and use a new stream for execution. More...

void	SetStream (cudaStream_t stream)
	Set the stream that the device is operating on. More...

const char *	GetPrototxtPath () const
	Retrieve the path to the network prototxt file. More...

const char *	GetModelPath () const
	Retrieve the full path to model file, including the filename. More...

const char *	GetModelFilename () const
	Retrieve the filename of the file, excluding the directory. More...

modelType	GetModelType () const
	Retrieve the format of the network model. More...

bool	IsModelType (modelType type) const
	Return true if the model is of the specified format. More...

uint32_t	GetInputLayers () const
	Retrieve the number of input layers to the network. More...

uint32_t	GetOutputLayers () const
	Retrieve the number of output layers to the network. More...

Dims3	GetInputDims (uint32_t layer=0) const
	Retrieve the dimensions of network input layer. More...

uint32_t	GetInputWidth (uint32_t layer=0) const
	Retrieve the width of network input layer. More...

uint32_t	GetInputHeight (uint32_t layer=0) const
	Retrieve the height of network input layer. More...

uint32_t	GetInputSize (uint32_t layer=0) const
	Retrieve the size (in bytes) of network input layer. More...

float *	GetInputPtr (uint32_t layer=0) const
	Get the CUDA pointer to the input layer's memory. More...

Dims3	GetOutputDims (uint32_t layer=0) const
	Retrieve the dimensions of network output layer. More...

uint32_t	GetOutputWidth (uint32_t layer=0) const
	Retrieve the width of network output layer. More...

uint32_t	GetOutputHeight (uint32_t layer=0) const
	Retrieve the height of network output layer. More...

uint32_t	GetOutputSize (uint32_t layer=0) const
	Retrieve the size (in bytes) of network output layer. More...

float *	GetOutputPtr (uint32_t layer=0) const
	Get the CUDA pointer to the output memory. More...

float	GetNetworkFPS ()
	Retrieve the network frames per second (FPS). More...

float	GetNetworkTime ()
	Retrieve the network runtime (in milliseconds). More...

const char *	GetNetworkName () const
	Retrieve the network name (it's filename). More...

float2	GetProfilerTime (profilerQuery query)
	Retrieve the profiler runtime (in milliseconds). More...

float	GetProfilerTime (profilerQuery query, profilerDevice device)
	Retrieve the profiler runtime (in milliseconds). More...

void	PrintProfilerTimes ()
	Print the profiler times (in millseconds). More...

Static Public Member Functions
static uint32_t	OverlayFlagsFromStr (const char *flags)
	Parse a string sequence into OverlayFlags enum. More...

static detectNet *	Create (const char *network="ssd-mobilenet-v2", float threshold=DETECTNET_DEFAULT_CONFIDENCE_THRESHOLD, uint32_t maxBatchSize=DEFAULT_MAX_BATCH_SIZE, precisionType precision=TYPE_FASTEST, deviceType device=DEVICE_GPU, bool allowGPUFallback=true)
	Load a pre-trained model. More...

static detectNet *	Create (const char prototxt_path, const char model_path, float mean_pixel=0.0f, const char class_labels=NULL, float threshold=DETECTNET_DEFAULT_CONFIDENCE_THRESHOLD, const char input=DETECTNET_DEFAULT_INPUT, const char coverage=DETECTNET_DEFAULT_COVERAGE, const char bboxes=DETECTNET_DEFAULT_BBOX, uint32_t maxBatchSize=DEFAULT_MAX_BATCH_SIZE, precisionType precision=TYPE_FASTEST, deviceType device=DEVICE_GPU, bool allowGPUFallback=true)
	Load a custom network instance. More...

static detectNet *	Create (const char prototxt_path, const char model_path, float mean_pixel, const char class_labels, const char class_colors, float threshold=DETECTNET_DEFAULT_CONFIDENCE_THRESHOLD, const char input=DETECTNET_DEFAULT_INPUT, const char coverage=DETECTNET_DEFAULT_COVERAGE, const char *bboxes=DETECTNET_DEFAULT_BBOX, uint32_t maxBatchSize=DEFAULT_MAX_BATCH_SIZE, precisionType precision=TYPE_FASTEST, deviceType device=DEVICE_GPU, bool allowGPUFallback=true)
	Load a custom network instance. More...

static detectNet *	Create (const char model_path, const char class_labels, float threshold, const char input, const Dims3 &inputDims, const char output, const char *numDetections, uint32_t maxBatchSize=DEFAULT_MAX_BATCH_SIZE, precisionType precision=TYPE_FASTEST, deviceType device=DEVICE_GPU, bool allowGPUFallback=true)
	Load a custom network instance of a UFF model. More...

static detectNet *	Create (int argc, char **argv)
	Load a new network instance by parsing the command line. More...

static detectNet *	Create (const commandLine &cmdLine)
	Load a new network instance by parsing the command line. More...

static const char *	Usage ()
	Usage string for command line arguments to Create() More...

Static Public Member Functions inherited from tensorNet
static bool	LoadClassLabels (const char *filename, std::vector< std::string > &descriptions, int expectedClasses=-1)
	Load class descriptions from a label file. More...

static bool	LoadClassLabels (const char *filename, std::vector< std::string > &descriptions, std::vector< std::string > &synsets, int expectedClasses=-1)
	Load class descriptions and synset strings from a label file. More...

static bool	LoadClassColors (const char filename, float4 colors, int expectedClasses, float defaultAlpha=255.0f)
	Load class colors from a text file. More...

static bool	LoadClassColors (const char filename, float4 *colors, int expectedClasses, float defaultAlpha=255.0f)
	Load class colors from a text file. More...

static float4	GenerateColor (uint32_t classID, float alpha=255.0f)
	Procedurally generate a color for a given class index with the specified alpha value. More...

static precisionType	SelectPrecision (precisionType precision, deviceType device=DEVICE_GPU, bool allowInt8=true)
	Resolve a desired precision to a specific one that's available. More...

static precisionType	FindFastestPrecision (deviceType device=DEVICE_GPU, bool allowInt8=true)
	Determine the fastest native precision on a device. More...

static std::vector< precisionType >	DetectNativePrecisions (deviceType device=DEVICE_GPU)
	Detect the precisions supported natively on a device. More...

static bool	DetectNativePrecision (const std::vector< precisionType > &nativeTypes, precisionType type)
	Detect if a particular precision is supported natively. More...

static bool	DetectNativePrecision (precisionType precision, deviceType device=DEVICE_GPU)
	Detect if a particular precision is supported natively. More...

Protected Member Functions
	detectNet (float meanPixel=0.0f)

bool	allocDetections ()

bool	loadClassInfo (const char *filename)

bool	loadClassColors (const char *filename)

bool	init (const char prototxt_path, const char model_path, const char class_labels, const char class_colors, float threshold, const char input, const char coverage, const char *bboxes, uint32_t maxBatchSize, precisionType precision, deviceType device, bool allowGPUFallback)

bool	preProcess (void *input, uint32_t width, uint32_t height, imageFormat format)

int	postProcess (void input, uint32_t width, uint32_t height, imageFormat format, Detection detections)

int	postProcessSSD_UFF (Detection *detections, uint32_t width, uint32_t height)

int	postProcessSSD_ONNX (Detection *detections, uint32_t width, uint32_t height)

int	postProcessDetectNet (Detection *detections, uint32_t width, uint32_t height)

int	postProcessDetectNet_v2 (Detection *detections, uint32_t width, uint32_t height)

int	clusterDetections (Detection *detections, int n)

void	sortDetections (Detection *detections, int numDetections)

Protected Member Functions inherited from tensorNet
	tensorNet ()
	Constructor. More...

bool	ProcessNetwork (bool sync=true)
	Execute processing of the network. More...

bool	ProfileModel (const std::string &deployFile, const std::string &modelFile, const std::vector< std::string > &inputs, const std::vector< Dims3 > &inputDims, const std::vector< std::string > &outputs, uint32_t maxBatchSize, precisionType precision, deviceType device, bool allowGPUFallback, nvinfer1::IInt8Calibrator calibrator, char engineStream, size_t engineSize)
	Create and output an optimized network model. More...

bool	ConfigureBuilder (nvinfer1::IBuilder builder, uint32_t maxBatchSize, uint32_t workspaceSize, precisionType precision, deviceType device, bool allowGPUFallback, nvinfer1::IInt8Calibrator calibrator)
	Configure builder options. More...

bool	ValidateEngine (const char model_path, const char cache_path, const char *checksum_path)
	Validate that the model already has a built TensorRT engine that exists and doesn't need updating. More...

void	PROFILER_BEGIN (profilerQuery query)
	Begin a profiling query, before network is run. More...

void	PROFILER_END (profilerQuery query)
	End a profiling query, after the network is run. More...

bool	PROFILER_QUERY (profilerQuery query)
	Query the CUDA part of a profiler query. More...

Protected Attributes
objectTracker *	mTracker

float	mConfidenceThreshold

float	mClusteringThreshold

float	mMeanPixel

float	mLineWidth

float	mOverlayAlpha

float4 *	mClassColors

std::vector< std::string >	mClassDesc

std::vector< std::string >	mClassSynset

std::string	mClassPath

uint32_t	mNumClasses

Detection *	mDetectionSets

uint32_t	mDetectionSet

uint32_t	mMaxDetections

Protected Attributes inherited from tensorNet
tensorNet::Logger	gLogger

tensorNet::Profiler	gProfiler

std::string	mPrototxtPath

std::string	mModelPath

std::string	mModelFile

std::string	mMeanPath

std::string	mCacheEnginePath

std::string	mCacheCalibrationPath

std::string	mChecksumPath

deviceType	mDevice

precisionType	mPrecision

modelType	mModelType

cudaStream_t	mStream

cudaEvent_t	mEventsGPU [PROFILER_TOTAL *2]

timespec	mEventsCPU [PROFILER_TOTAL *2]

nvinfer1::IRuntime *	mInfer

nvinfer1::ICudaEngine *	mEngine

nvinfer1::IExecutionContext *	mContext

float2	mProfilerTimes [PROFILER_TOTAL+1]

uint32_t	mProfilerQueriesUsed

uint32_t	mProfilerQueriesDone

uint32_t	mWorkspaceSize

uint32_t	mMaxBatchSize

bool	mEnableProfiler

bool	mEnableDebug

bool	mAllowGPUFallback

void **	mBindings

std::vector< layerInfo >	mInputs

std::vector< layerInfo >	mOutputs

Static Protected Attributes
static const uint32_t	mNumDetectionSets = 16

Member Enumeration Documentation

◆ OverlayFlags

enum detectNet::OverlayFlags

Overlay flags (can be OR'd together).

Enumerator
OVERLAY_NONE	No overlay.
OVERLAY_BOX	Overlay the object bounding boxes (filled)
OVERLAY_LABEL	Overlay the class description labels.
OVERLAY_CONFIDENCE	Overlay the detection confidence values.
OVERLAY_TRACKING	Overlay tracking information (like track ID)
OVERLAY_LINES	Overlay the bounding box lines (unfilled)
OVERLAY_DEFAULT	The default choice of overlay.

Constructor & Destructor Documentation

◆ ~detectNet()

virtual detectNet::~detectNet ( )

virtual

Destory.

◆ detectNet()

detectNet::detectNet ( float meanPixel = 0.0f )

protected

Member Function Documentation

◆ allocDetections()

bool detectNet::allocDetections ( )

protected

◆ clusterDetections()

int detectNet::clusterDetections	(	Detection *	detections,
		int	n
	)

protected

◆ Create() [1/6]

static detectNet* detectNet::Create	(	const char *	model_path,
		const char *	class_labels,
		float	threshold,
		const char *	input,
		const Dims3 &	inputDims,
		const char *	output,
		const char *	numDetections,
		uint32_t	maxBatchSize = `DEFAULT_MAX_BATCH_SIZE`,
		precisionType	precision = `TYPE_FASTEST`,
		deviceType	device = `DEVICE_GPU`,
		bool	allowGPUFallback = `true`
	)

static

Load a custom network instance of a UFF model.

Parameters

model_path	File path to the UFF model
class_labels	File path to list of class name labels
threshold	default minimum threshold for detection
input	Name of the input layer blob.
inputDims	Dimensions of the input layer blob.
output	Name of the output layer blob containing the bounding boxes, ect.
numDetections	Name of the output layer blob containing the detection count.
maxBatchSize	The maximum batch size that the network will support and be optimized for.

◆ Create() [2/6]

static detectNet* detectNet::Create	(	const char *	network = `"ssd-mobilenet-v2"`,
		float	threshold = `DETECTNET_DEFAULT_CONFIDENCE_THRESHOLD`,
		uint32_t	maxBatchSize = `DEFAULT_MAX_BATCH_SIZE`,
		precisionType	precision = `TYPE_FASTEST`,
		deviceType	device = `DEVICE_GPU`,
		bool	allowGPUFallback = `true`
	)

static

Load a pre-trained model.

Parameters

network the pre-trained model to load (

See also: DETECTNET_USAGE_STRING for models)

Parameters

threshold	default minimum threshold for detection
maxBatchSize	The maximum batch size that the network will support and be optimized for.

◆ Create() [3/6]

static detectNet* detectNet::Create	(	const char *	prototxt_path,
		const char *	model_path,
		float	mean_pixel,
		const char *	class_labels,
		const char *	class_colors,
		float	threshold = `DETECTNET_DEFAULT_CONFIDENCE_THRESHOLD`,
		const char *	input = `DETECTNET_DEFAULT_INPUT`,
		const char *	coverage = `DETECTNET_DEFAULT_COVERAGE`,
		const char *	bboxes = `DETECTNET_DEFAULT_BBOX`,
		uint32_t	maxBatchSize = `DEFAULT_MAX_BATCH_SIZE`,
		precisionType	precision = `TYPE_FASTEST`,
		deviceType	device = `DEVICE_GPU`,
		bool	allowGPUFallback = `true`
	)

static

Load a custom network instance.

Parameters

prototxt_path	File path to the deployable network prototxt
model_path	File path to the caffemodel
mean_pixel	Input transform subtraction value (use 0.0 if the network already does this)
class_labels	File path to list of class name labels
class_colors	File path to list of class colors
threshold	default minimum threshold for detection
input	Name of the input layer blob.
coverage	Name of the output coverage classifier layer blob, which contains the confidence values for each bbox.
bboxes	Name of the output bounding box layer blob, which contains a grid of rectangles in the image.
maxBatchSize	The maximum batch size that the network will support and be optimized for.

◆ Create() [4/6]

static detectNet* detectNet::Create	(	const char *	prototxt_path,
		const char *	model_path,
		float	mean_pixel = `0.0f`,
		const char *	class_labels = `NULL`,
		float	threshold = `DETECTNET_DEFAULT_CONFIDENCE_THRESHOLD`,
		const char *	input = `DETECTNET_DEFAULT_INPUT`,
		const char *	coverage = `DETECTNET_DEFAULT_COVERAGE`,
		const char *	bboxes = `DETECTNET_DEFAULT_BBOX`,
		uint32_t	maxBatchSize = `DEFAULT_MAX_BATCH_SIZE`,
		precisionType	precision = `TYPE_FASTEST`,
		deviceType	device = `DEVICE_GPU`,
		bool	allowGPUFallback = `true`
	)

static

Load a custom network instance.

Parameters

prototxt_path	File path to the deployable network prototxt
model_path	File path to the caffemodel
mean_pixel	Input transform subtraction value (use 0.0 if the network already does this)
class_labels	File path to list of class name labels
threshold	default minimum threshold for detection
input	Name of the input layer blob.
coverage	Name of the output coverage classifier layer blob, which contains the confidence values for each bbox.
bboxes	Name of the output bounding box layer blob, which contains a grid of rectangles in the image.
maxBatchSize	The maximum batch size that the network will support and be optimized for.

◆ Create() [5/6]

static detectNet* detectNet::Create ( const commandLine & cmdLine )

static

Load a new network instance by parsing the command line.

◆ Create() [6/6]

static detectNet* detectNet::Create	(	int	argc,
		char **	argv
	)

static

Load a new network instance by parsing the command line.

◆ Detect() [1/6]

int detectNet::Detect	(	float *	input,
		uint32_t	width,
		uint32_t	height,
		Detection **	detections,
		uint32_t	overlay = `OVERLAY_DEFAULT`
	)

Detect object locations from an RGBA image, returning an array containing the detection results.

Deprecated:: this overload of Detect() provides legacy compatibility with float* type (RGBA32F).

Parameters

[in]	input	float4 RGBA input image in CUDA device memory.
[in]	width	width of the input image in pixels.
[in]	height	height of the input image in pixels.
[out]	detections	pointer that will be set to array of detection results (residing in shared CPU/GPU memory)
[in]	overlay	bitwise OR combination of overlay flags (

See also: OverlayFlags and; Overlay()), or OVERLAY_NONE.

Returns: The number of detected objects, 0 if there were no detected objects, and -1 if an error was encountered.

◆ Detect() [2/6]

int detectNet::Detect	(	float *	input,
		uint32_t	width,
		uint32_t	height,
		Detection *	detections,
		uint32_t	overlay = `OVERLAY_DEFAULT`
	)

Detect object locations in an RGBA image, into an array of the results allocated by the user.

Deprecated:: this overload of Detect() provides legacy compatibility with float* type (RGBA32F).

Parameters

[in]	input	float4 RGBA input image in CUDA device memory.
[in]	width	width of the input image in pixels.
[in]	height	height of the input image in pixels.
[out]	detections	pointer to user-allocated array that will be filled with the detection results.

See also: GetMaxDetections() for the number of detection results that should be allocated in this buffer.

Parameters

[in] overlay bitwise OR combination of overlay flags (

See also: OverlayFlags and; Overlay()), or OVERLAY_NONE.

Returns: The number of detected objects, 0 if there were no detected objects, and -1 if an error was encountered.

◆ Detect() [3/6]

template<typename T >

int detectNet::Detect	(	T *	image,
		uint32_t	width,
		uint32_t	height,
		Detection **	detections,
		uint32_t	overlay = `OVERLAY_DEFAULT`
	)

inline

Detect object locations from an image, returning an array containing the detection results.

Parameters

[in]	input	input image in CUDA device memory (uchar3/uchar4/float3/float4)
[in]	width	width of the input image in pixels.
[in]	height	height of the input image in pixels.
[out]	detections	pointer that will be set to array of detection results (residing in shared CPU/GPU memory)
[in]	overlay	bitwise OR combination of overlay flags (

See also: OverlayFlags and; Overlay()), or OVERLAY_NONE.

Returns: The number of detected objects, 0 if there were no detected objects, and -1 if an error was encountered.

◆ Detect() [4/6]

template<typename T >

int detectNet::Detect	(	T *	image,
		uint32_t	width,
		uint32_t	height,
		Detection *	detections,
		uint32_t	overlay = `OVERLAY_DEFAULT`
	)

inline

Detect object locations in an image, into an array of the results allocated by the user.

Parameters

[in]	input	input image in CUDA device memory (uchar3/uchar4/float3/float4)
[in]	width	width of the input image in pixels.
[in]	height	height of the input image in pixels.
[out]	detections	pointer to user-allocated array that will be filled with the detection results.

See also: GetMaxDetections() for the number of detection results that should be allocated in this buffer.

Parameters

[in] overlay bitwise OR combination of overlay flags (

See also: OverlayFlags and; Overlay()), or OVERLAY_NONE.

Returns: The number of detected objects, 0 if there were no detected objects, and -1 if an error was encountered.

◆ Detect() [5/6]

int detectNet::Detect	(	void *	input,
		uint32_t	width,
		uint32_t	height,
		imageFormat	format,
		Detection **	detections,
		uint32_t	overlay = `OVERLAY_DEFAULT`
	)

Detect object locations from an image, returning an array containing the detection results.

Parameters

[in]	input	input image in CUDA device memory (uchar3/uchar4/float3/float4)
[in]	width	width of the input image in pixels.
[in]	height	height of the input image in pixels.
[out]	detections	pointer that will be set to array of detection results (residing in shared CPU/GPU memory)
[in]	overlay	bitwise OR combination of overlay flags (

See also: OverlayFlags and; Overlay()), or OVERLAY_NONE.

Returns: The number of detected objects, 0 if there were no detected objects, and -1 if an error was encountered.

◆ Detect() [6/6]

int detectNet::Detect	(	void *	input,
		uint32_t	width,
		uint32_t	height,
		imageFormat	format,
		Detection *	detections,
		uint32_t	overlay = `OVERLAY_DEFAULT`
	)

Detect object locations from an image, into an array of the results allocated by the user.

Parameters

[in]	input	input image in CUDA device memory (uchar3/uchar4/float3/float4)
[in]	width	width of the input image in pixels.
[in]	height	height of the input image in pixels.
[out]	detections	pointer to user-allocated array that will be filled with the detection results.

See also: GetMaxDetections() for the number of detection results that should be allocated in this buffer.

Parameters

[in] overlay bitwise OR combination of overlay flags (

See also: OverlayFlags and; Overlay()), or OVERLAY_NONE.

Returns: The number of detected objects, 0 if there were no detected objects, and -1 if an error was encountered.

◆ GetClassColor()

float4 detectNet::GetClassColor ( uint32_t classIndex ) const

inline

Retrieve the RGBA visualization color a particular class.

◆ GetClassDesc()

const char* detectNet::GetClassDesc ( uint32_t index ) const

inline

Retrieve the description of a particular class.

◆ GetClassLabel()

const char* detectNet::GetClassLabel ( uint32_t index ) const

inline

Retrieve the description of a particular class.

◆ GetClassPath()

const char* detectNet::GetClassPath ( ) const

inline

Retrieve the path to the file containing the class descriptions.

◆ GetClassSynset()

const char* detectNet::GetClassSynset ( uint32_t index ) const

inline

Retrieve the class synset category of a particular class.

◆ GetClusteringThreshold()

float detectNet::GetClusteringThreshold ( ) const

inline

Retrieve the overlapping area % threshold for clustering.

◆ GetConfidenceThreshold()

float detectNet::GetConfidenceThreshold ( ) const

inline

Retrieve the minimum threshold for detection.

◆ GetLineWidth()

float detectNet::GetLineWidth ( ) const

inline

Retrieve the line width used during overlay when OVERLAY_LINES is used.

◆ GetMaxDetections()

uint32_t detectNet::GetMaxDetections ( ) const

inline

Retrieve the maximum number of simultaneous detections the network supports.

Knowing this is useful for allocating the buffers to store the output detection results.

◆ GetNumClasses()

uint32_t detectNet::GetNumClasses ( ) const

inline

Retrieve the number of object classes supported in the detector.

◆ GetOverlayAlpha()

float detectNet::GetOverlayAlpha ( ) const

inline

Retrieve the overlay alpha blending value for classes that don't have it explicitly set (between 0-255).

◆ GetThreshold()

float detectNet::GetThreshold ( ) const

inline

Retrieve the minimum threshold for detection.

Deprecated:: please use GetConfidenceThreshold() instead

◆ GetTracker()

objectTracker* detectNet::GetTracker ( ) const

inline

Get the object tracker being used.

◆ init()

bool detectNet::init	(	const char *	prototxt_path,
		const char *	model_path,
		const char *	class_labels,
		const char *	class_colors,
		float	threshold,
		const char *	input,
		const char *	coverage,
		const char *	bboxes,
		uint32_t	maxBatchSize,
		precisionType	precision,
		deviceType	device,
		bool	allowGPUFallback
	)

protected

◆ loadClassColors()

bool detectNet::loadClassColors ( const char * filename )

protected

◆ loadClassInfo()

bool detectNet::loadClassInfo ( const char * filename )

protected

◆ Overlay() [1/2]

template<typename T >

bool detectNet::Overlay	(	T *	input,
		T *	output,
		uint32_t	width,
		uint32_t	height,
		Detection *	detections,
		uint32_t	numDetections,
		uint32_t	flags = `OVERLAY_DEFAULT`
	)

inline

Draw the detected bounding boxes overlayed on an RGBA image.

Note: Overlay() will automatically be called by default by Detect(), if the overlay parameter is true

Parameters

input	input image in CUDA device memory.
output	output image in CUDA device memory.
detections	Array of detections allocated in CUDA device memory.

◆ Overlay() [2/2]

bool detectNet::Overlay	(	void *	input,
		void *	output,
		uint32_t	width,
		uint32_t	height,
		imageFormat	format,
		Detection *	detections,
		uint32_t	numDetections,
		uint32_t	flags = `OVERLAY_DEFAULT`
	)

Draw the detected bounding boxes overlayed on an RGBA image.

Note: Overlay() will automatically be called by default by Detect(), if the overlay parameter is true

Parameters

input	input image in CUDA device memory.
output	output image in CUDA device memory.
detections	Array of detections allocated in CUDA device memory.

◆ OverlayFlagsFromStr()

static uint32_t detectNet::OverlayFlagsFromStr ( const char * flags )

static

Parse a string sequence into OverlayFlags enum.

Valid flags are "none", "box", "label", and "conf" and it is possible to combine flags (bitwise OR) together with commas or pipe (|) symbol. For example, the string sequence "box,label,conf" would return the flags OVERLAY_BOX|OVERLAY_LABEL|OVERLAY_CONFIDENCE.

◆ postProcess()

int detectNet::postProcess	(	void *	input,
		uint32_t	width,
		uint32_t	height,
		imageFormat	format,
		Detection *	detections
	)

protected

◆ postProcessDetectNet()

int detectNet::postProcessDetectNet	(	Detection *	detections,
		uint32_t	width,
		uint32_t	height
	)

protected

◆ postProcessDetectNet_v2()

int detectNet::postProcessDetectNet_v2	(	Detection *	detections,
		uint32_t	width,
		uint32_t	height
	)

protected

◆ postProcessSSD_ONNX()

int detectNet::postProcessSSD_ONNX	(	Detection *	detections,
		uint32_t	width,
		uint32_t	height
	)

protected

◆ postProcessSSD_UFF()

int detectNet::postProcessSSD_UFF	(	Detection *	detections,
		uint32_t	width,
		uint32_t	height
	)

protected

◆ preProcess()

bool detectNet::preProcess	(	void *	input,
		uint32_t	width,
		uint32_t	height,
		imageFormat	format
	)

protected

◆ SetClassColor() [1/2]

void detectNet::SetClassColor	(	uint32_t	classIndex,
		const float4 &	color
	)

inline

Set the visualization color of a particular class of object.

◆ SetClassColor() [2/2]

void detectNet::SetClassColor	(	uint32_t	classIndex,
		float	r,
		float	g,
		float	b,
		float	a = `255.0f`
	)

inline

Set the visualization color of a particular class of object.

◆ SetClusteringThreshold()

void detectNet::SetClusteringThreshold ( float threshold )

inline

Set the overlapping area % threshold for clustering.

◆ SetConfidenceThreshold()

void detectNet::SetConfidenceThreshold ( float threshold )

inline

Set the minimum threshold for detection.

◆ SetLineWidth()

void detectNet::SetLineWidth ( float width )

inline

Set the line width used during overlay when OVERLAY_LINES is used.

◆ SetOverlayAlpha()

void detectNet::SetOverlayAlpha ( float alpha )

Set overlay alpha blending value for all classes (between 0-255).

◆ SetThreshold()

void detectNet::SetThreshold ( float threshold )

inline

Set the minimum threshold for detection.

Deprecated:: please use SetConfidenceThreshold() instead

◆ SetTracker()

void detectNet::SetTracker ( objectTracker * tracker )

inline

Set the object tracker to be used.

◆ sortDetections()

void detectNet::sortDetections	(	Detection *	detections,
		int	numDetections
	)

protected

◆ Usage()

static const char* detectNet::Usage ( )

inlinestatic

Usage string for command line arguments to Create()

Member Data Documentation

◆ mClassColors

float4* detectNet::mClassColors

protected

◆ mClassDesc

std::vector<std::string> detectNet::mClassDesc

protected

◆ mClassPath

std::string detectNet::mClassPath

protected

◆ mClassSynset

std::vector<std::string> detectNet::mClassSynset

protected

◆ mClusteringThreshold

float detectNet::mClusteringThreshold

protected

◆ mConfidenceThreshold

float detectNet::mConfidenceThreshold

protected

◆ mDetectionSet

uint32_t detectNet::mDetectionSet

protected

◆ mDetectionSets

Detection* detectNet::mDetectionSets

protected

◆ mLineWidth

float detectNet::mLineWidth

protected

◆ mMaxDetections

uint32_t detectNet::mMaxDetections

protected

◆ mMeanPixel

float detectNet::mMeanPixel

protected

◆ mNumClasses

uint32_t detectNet::mNumClasses

protected

◆ mNumDetectionSets

const uint32_t detectNet::mNumDetectionSets = 16

staticprotected

◆ mOverlayAlpha

float detectNet::mOverlayAlpha

protected

◆ mTracker

objectTracker* detectNet::mTracker

protected

Macro Definition Documentation

◆ DETECTNET_DEFAULT_ALPHA

#define DETECTNET_DEFAULT_ALPHA 120

Default alpha blending value used during overlay.

◆ DETECTNET_DEFAULT_BBOX

#define DETECTNET_DEFAULT_BBOX "bboxes"

Name of default output blob of the grid of bounding boxes for DetectNet caffe model.

◆ DETECTNET_DEFAULT_CLUSTERING_THRESHOLD

#define DETECTNET_DEFAULT_CLUSTERING_THRESHOLD 0.75f

Default value of the clustering area-of-overlap threshold.

◆ DETECTNET_DEFAULT_CONFIDENCE_THRESHOLD

#define DETECTNET_DEFAULT_CONFIDENCE_THRESHOLD 0.5f

Default value of the minimum detection threshold.

◆ DETECTNET_DEFAULT_COVERAGE

#define DETECTNET_DEFAULT_COVERAGE "coverage"

Name of default output blob of the coverage map for DetectNet caffe model.

◆ DETECTNET_DEFAULT_INPUT

#define DETECTNET_DEFAULT_INPUT "data"

Name of default input blob for DetectNet caffe model.

◆ DETECTNET_DEFAULT_THRESHOLD

#define DETECTNET_DEFAULT_THRESHOLD DETECTNET_DEFAULT_CONFIDENCE_THRESHOLD

Default value of the minimum detection threshold.

Deprecated:: please use DETECTNET_DEFAULT_CONFIDENCE_THRESHOLD instead

◆ DETECTNET_MODEL_TYPE

#define DETECTNET_MODEL_TYPE "detection"

The model type for detectNet in data/networks/models.json.

◆ DETECTNET_USAGE_STRING

#define DETECTNET_USAGE_STRING

Value:

                  "detectNet arguments: \n"                                     \
                  "  --network=NETWORK     pre-trained model to load, one of the following:\n"          \
                  "                            * ssd-mobilenet-v1\n"                                                    \
                  "                            * ssd-mobilenet-v2 (default)\n"                                  \
                  "                            * ssd-inception-v2\n"                                                    \
                  "                            * peoplenet\n"                                        \
                  "                            * peoplenet-pruned\n"                                 \
                  "                            * dashcamnet\n"                                       \
                  "                            * trafficcamnet\n"                                    \
                  "                            * facedetect\n"                                       \
                  "  --model=MODEL         path to custom model to load (caffemodel, uff, or onnx)\n"                                   \
                  "  --prototxt=PROTOTXT   path to custom prototxt to load (for .caffemodel only)\n"                                    \
                  "  --labels=LABELS       path to text file containing the labels for each class\n"                                    \
                  "  --input-blob=INPUT    name of the input layer (default is '" DETECTNET_DEFAULT_INPUT "')\n"                        \
                  "  --output-cvg=COVERAGE name of the coverage/confidence output layer (default is '" DETECTNET_DEFAULT_COVERAGE "')\n"        \
                  "  --output-bbox=BOXES   name of the bounding output layer (default is '" DETECTNET_DEFAULT_BBOX "')\n"       \
                  "  --mean-pixel=PIXEL    mean pixel value to subtract from input (default is 0.0)\n"                                  \
                  "  --confidence=CONF     minimum confidence threshold for detection (default is 0.5)\n"                               \
                  "  --clustering=CLUSTER  minimum overlapping area threshold for clustering (default is 0.75)\n"             \
            "  --alpha=ALPHA         overlay alpha blending value, range 0-255 (default: 120)\n"                                        \
                  "  --overlay=OVERLAY     detection overlay flags (e.g. --overlay=box,labels,conf)\n"                                  \
                  "                        valid combinations are:  'box', 'lines', 'labels', 'conf', 'none'\n"                 \
                  "  --profile             enable layer profiling in TensorRT\n\n"                              \

Standard command-line options able to be passed to detectNet::Create()

Classes

Macros

Detailed Description

Class Documentation

◆ detectNet

Public Types

Public Member Functions

Static Public Member Functions

Protected Member Functions

Protected Attributes

Static Protected Attributes

Member Enumeration Documentation

◆ OverlayFlags

Constructor & Destructor Documentation

◆ ~detectNet()

◆ detectNet()

Member Function Documentation

◆ allocDetections()

◆ clusterDetections()

◆ Create() [1/6]

◆ Create() [2/6]

◆ Create() [3/6]

◆ Create() [4/6]

◆ Create() [5/6]

◆ Create() [6/6]

◆ Detect() [1/6]

◆ Detect() [2/6]

◆ Detect() [3/6]

◆ Detect() [4/6]

◆ Detect() [5/6]

◆ Detect() [6/6]

◆ GetClassColor()

◆ GetClassDesc()

◆ GetClassLabel()

◆ GetClassPath()

◆ GetClassSynset()

◆ GetClusteringThreshold()

◆ GetConfidenceThreshold()

◆ GetLineWidth()

◆ GetMaxDetections()

◆ GetNumClasses()

◆ GetOverlayAlpha()

◆ GetThreshold()

◆ GetTracker()

◆ init()

◆ loadClassColors()

◆ loadClassInfo()

◆ Overlay() [1/2]

◆ Overlay() [2/2]

◆ OverlayFlagsFromStr()

◆ postProcess()

◆ postProcessDetectNet()

◆ postProcessDetectNet_v2()

◆ postProcessSSD_ONNX()

◆ postProcessSSD_UFF()

◆ preProcess()

◆ SetClassColor() [1/2]

◆ SetClassColor() [2/2]

◆ SetClusteringThreshold()

◆ SetConfidenceThreshold()

◆ SetLineWidth()

◆ SetOverlayAlpha()

◆ SetThreshold()

◆ SetTracker()

◆ sortDetections()

◆ Usage()

Member Data Documentation

◆ mClassColors

◆ mClassDesc

◆ mClassPath

◆ mClassSynset

◆ mClusteringThreshold

◆ mConfidenceThreshold

◆ mDetectionSet

◆ mDetectionSets

◆ mLineWidth

◆ mMaxDetections

◆ mMeanPixel

◆ mNumClasses

◆ mNumDetectionSets