Structure containing a set of parameters for the object detection module. More...

Functions
	ObjectDetectionParameters (bool enable_tracking_=true, bool enable_segmentation_=false, OBJECT_DETECTION_MODEL detection_model=OBJECT_DETECTION_MODEL::MULTI_CLASS_BOX_FAST, float max_range_=-1.f, BatchParameters batch_trajectories_parameters=BatchParameters(), OBJECT_FILTERING_MODE filtering_mode_=OBJECT_FILTERING_MODE::NMS3D, float prediction_timeout_s=0.2f, bool allow_reduced_precision_inference=false, unsigned int instance_id=0, const sl::String &fused_objects_group_name="", const sl::String &custom_onnx_file="", const sl::Resolution &custom_onnx_dynamic_input_shape=sl::Resolution(512, 512))
	Default constructor. More...

bool	save (String filename) const
	Saves the current set of parameters into a file to be reloaded with the load() method. More...

bool	load (String filename)
	Loads a set of parameters from the values contained in a previously saved file. More...

bool	encode (String &serialized_content) const
	Generate a JSON Object (with the struct type as a key) containing the serialized struct, converted into a string. More...

bool	decode (const String &serialized_content)
	Fill the structure from the serialized json object contained in the input string. More...

bool	operator== (const ObjectDetectionParameters &param1) const

bool	operator!= (const ObjectDetectionParameters &param1) const

Attributes
unsigned int	instance_module_id = 0
	Id of the module instance. More...

bool	enable_tracking = true
	Whether the object detection system includes object tracking capabilities across a sequence of images. More...

bool	enable_segmentation = false
	Whether the object masks will be computed. More...

OBJECT_DETECTION_MODEL	detection_model = OBJECT_DETECTION_MODEL::MULTI_CLASS_BOX_FAST
	sl::OBJECT_DETECTION_MODEL to use. More...

sl::String	fused_objects_group_name
	In a multi camera setup, specify which group this model belongs to. More...

sl::String	custom_onnx_file
	Path to the YOLO-like onnx file for custom object detection ran in the ZED SDK. More...

sl::Resolution	custom_onnx_dynamic_input_shape
	Resolution to the YOLO-like onnx file for custom object detection ran in the ZED SDK. This resolution defines the input tensor size for dynamic shape ONNX model only. The batch and channel dimensions are automatically handled, it assumes it's color images like default YOLO models. More...

float	max_range = -1.f
	Upper depth range for detections. More...

BatchParameters	batch_parameters
	Batching system parameters. More...

OBJECT_FILTERING_MODE	filtering_mode = OBJECT_FILTERING_MODE::NMS3D
	Filtering mode that should be applied to raw detections. More...

float	prediction_timeout_s
	Prediction duration of the ZED SDK when an object is not detected anymore before switching its state to sl::OBJECT_TRACKING_STATE::SEARCHING. More...

bool	allow_reduced_precision_inference
	Whether to allow inference to run at a lower precision to improve runtime and memory usage. More...

Detailed Description

Structure containing a set of parameters for the object detection module.

The default constructor sets all parameters to their default settings.

Note: Parameters can be adjusted by the user.

Constructor and Destructor

◆ ObjectDetectionParameters()

ObjectDetectionParameters	(	bool	enable_tracking_ = `true`,
		bool	enable_segmentation_ = `false`,
		OBJECT_DETECTION_MODEL	detection_model = `OBJECT_DETECTION_MODEL::MULTI_CLASS_BOX_FAST`,
		float	max_range_ = `-1.f`,
		BatchParameters	batch_trajectories_parameters = `BatchParameters()`,
		OBJECT_FILTERING_MODE	filtering_mode_ = `OBJECT_FILTERING_MODE::NMS3D`,
		float	prediction_timeout_s = `0.2f`,
		bool	allow_reduced_precision_inference = `false`,
		unsigned int	instance_id = `0`,
		const sl::String &	fused_objects_group_name = `""`,
		const sl::String &	custom_onnx_file = `""`,
		const sl::Resolution &	custom_onnx_dynamic_input_shape = `sl::Resolution(512, 512)`
	)

Default constructor.

All the parameters are set to their default values.

Functions

◆ save()

bool save ( String filename ) const

Saves the current set of parameters into a file to be reloaded with the load() method.

Parameters

filename : Name of the file which will be created to store the parameters (extension '.json' will be added if not set).

Returns: True if the file was successfully saved, otherwise false.

Warning: For security reasons, the file must not already exist.; In case a file already exists, the method will return false and existing file will not be updated.

◆ load()

bool load ( String filename )

Loads a set of parameters from the values contained in a previously saved file.

Parameters

filename : Path to the file from which the parameters will be loaded (extension '.json' will be added at the end of the filename if not detected).

Returns: True if the file was successfully loaded, otherwise false.

◆ encode()

bool encode ( String & serialized_content ) const

Generate a JSON Object (with the struct type as a key) containing the serialized struct, converted into a string.

Parameters

serialized_content output string containing the JSON Object

Returns: True if file was successfully saved, otherwise false.

◆ decode()

bool decode ( const String & serialized_content )

Fill the structure from the serialized json object contained in the input string.

Parameters

serialized_content input string containing the JSON Object

Returns: True if the decoding was successful, otherwise false.

◆ operator==()

bool operator== ( const ObjectDetectionParameters & param1 ) const

Comparison operator ==

Parameters

param1 to compare

Returns: true if the two struct are identical

◆ operator!=()

bool operator!= ( const ObjectDetectionParameters & param1 ) const

Comparison operator !=

Parameters

param1 to compare

Returns: true if the two struct are different

Variables

◆ instance_module_id

unsigned int instance_module_id = 0

Id of the module instance.

This is used to identify which object detection module instance is used.

◆ enable_tracking

bool enable_tracking = true

Whether the object detection system includes object tracking capabilities across a sequence of images.

◆ enable_segmentation

bool enable_segmentation = false

Whether the object masks will be computed.

◆ detection_model

OBJECT_DETECTION_MODEL detection_model = OBJECT_DETECTION_MODEL::MULTI_CLASS_BOX_FAST

sl::OBJECT_DETECTION_MODEL to use.

◆ fused_objects_group_name

sl::String fused_objects_group_name

In a multi camera setup, specify which group this model belongs to.

In a multi camera setup, multiple cameras can be used to detect objects and multiple detector having similar output layout can see the same object. Therefore, Fusion will fuse together the outputs received by multiple detectors only if they are part of the same fused_objects_group_name.

Note: This parameter is not used when not using a multi-camera setup and must be set in a multi camera setup.

◆ custom_onnx_file

sl::String custom_onnx_file

Path to the YOLO-like onnx file for custom object detection ran in the ZED SDK.

When detection_model is OBJECT_DETECTION_MODEL::CUSTOM_YOLOLIKE_BOX_OBJECTS, a onnx model must be passed so that the ZED SDK can optimize it for your GPU and run inference on it.

The resulting optimized model will be saved for re-use in the future.

Attention: - The model must be a YOLO-like model.; - The caching uses the deserialized custom_onnx_file along with your GPU specs to decide whether to use the cached optimized model or to optimize the passed onnx model. If you change the weights of the onnx file and pass the same path, the ZED SDK will detect the difference and optimize the new model.

Note: This parameter is useless when detection_model is not OBJECT_DETECTION_MODEL::CUSTOM_YOLOLIKE_BOX_OBJECTS.

◆ custom_onnx_dynamic_input_shape

sl::Resolution custom_onnx_dynamic_input_shape

Resolution to the YOLO-like onnx file for custom object detection ran in the ZED SDK. This resolution defines the input tensor size for dynamic shape ONNX model only. The batch and channel dimensions are automatically handled, it assumes it's color images like default YOLO models.

Note: This parameter is only used when detection_model is OBJECT_DETECTION_MODEL::CUSTOM_YOLOLIKE_BOX_OBJECTS and the provided ONNX file is using dynamic shapes.

Attention: - Multiple model only support squared images

\default Squared images 512x512 (input tensor will be 1x3x512x512)

◆ max_range

float max_range = -1.f

Upper depth range for detections.

Default: -1.f (value set in sl::InitParameters.depth_maximum_distance)

Note: The value cannot be greater than sl::InitParameters.depth_maximum_distance and its unit is defined in sl::InitParameters.coordinate_units.

◆ batch_parameters

BatchParameters batch_parameters

Batching system parameters.

Batching system (introduced in 3.5) performs short-term re-identification with deep-learning and trajectories filtering.
sl::BatchParameters.enable must to be true to use this feature (by default disabled).

◆ filtering_mode

OBJECT_FILTERING_MODE filtering_mode = OBJECT_FILTERING_MODE::NMS3D

Filtering mode that should be applied to raw detections.

Default: sl::OBJECT_FILTERING_MODE::NMS_3D (same behavior as previous ZED SDK version)

Note: This parameter is only used in detection model sl::OBJECT_DETECTION_MODEL::MULTI_CLASS_BOX_XXX and sl::OBJECT_DETECTION_MODEL::CUSTOM_BOX_OBJECTS.; For custom object, it is recommended to use sl::OBJECT_FILTERING_MODE::NMS_3D_PER_CLASS or sl::OBJECT_FILTERING_MODE::NONE.; In this case, you might need to add your own NMS filter before ingesting the boxes into the object detection module.

◆ prediction_timeout_s

float prediction_timeout_s

Prediction duration of the ZED SDK when an object is not detected anymore before switching its state to sl::OBJECT_TRACKING_STATE::SEARCHING.

It prevents the jittering of the object state when there is a short misdetection.
The user can define their own prediction time duration.

Note: During this time, the object will have sl::OBJECT_TRACKING_STATE::OK state even if it is not detected.; The duration is expressed in seconds.

Warning: prediction_timeout_s will be clamped to 1 second as the prediction is getting worse with time.; Setting this parameter to 0 disables the ZED SDK predictions.

◆ allow_reduced_precision_inference

bool allow_reduced_precision_inference

Whether to allow inference to run at a lower precision to improve runtime and memory usage.

It might increase the initial optimization time and could include downloading calibration data or calibration cache and slightly reduce the accuracy.

Note: The fp16 is automatically enabled if the GPU is compatible and provides a speed up of almost x2 and reduce memory usage by almost half, no precision loss.; This setting allow int8 precision which can speed up by another x2 factor (compared to fp16, or x4 compared to fp32) and half the fp16 memory usage, however some accuracy could be lost.; The accuracy loss should not exceed 1-2% on the compatible models.; The current compatible models are all sl::AI_MODELS::HUMAN_BODY_XXXX.

Functions

Attributes

Detailed Description

Constructor and Destructor

◆ ObjectDetectionParameters()

Functions

◆ save()

◆ load()

◆ encode()

◆ decode()

◆ operator==()

◆ operator!=()

Variables

◆ instance_module_id

◆ enable_tracking

◆ enable_segmentation

◆ detection_model

◆ fused_objects_group_name

◆ custom_onnx_file

◆ custom_onnx_dynamic_input_shape

◆ max_range

◆ batch_parameters

◆ filtering_mode

◆ prediction_timeout_s

◆ allow_reduced_precision_inference