Advanced modules samples

Advanced modules C++ samples

C++ samples of the advanced modules are showing some algorithms in action.

In the following table, for each sample, you will find:

• its module

• if the sample is using the SDK pipelines

• what UI framework is used (SDK UI module, OpenCV or ImGui)

• the main algorithms or classes demonstrated in the sample

Name	Description	Module	Using SDK pipelines	UI framework used for display	Main algorithms or classes demonstrated in the sample
metavision_blinking_pattern_focus	Focuses an event-based camera using a blinking pattern	Calibration	No	SDK UI	DftHighFreqScorerAlgorithm
metavision_mono_calibration	Estimates intrinsics parameters of an event-based camera using a blinking pattern	Calibration	Yes	OpenCV	BlinkingFrameGeneratorAlgorithm BlinkingDotsGridDetectorAlgorithm
metavision_noise_filtering	Filters events with a noise filter Activity, Trail or STC) and displays them on the screen	CV	No	SDK UI	ActivityNoiseFilterAlgorithm TrailFilterAlgorithm SpatioTemporalContrastAlgorithm
metavision_data_rate	Displays event rate from an event-based camera or from a recorded file	CV	No	SDK UI	ActivityNoiseFilterAlgorithm SpatioTemporalContrastAlgorithm AntiFlickerAlgorithm
metavision_sparse_optical_flow	Computes the sparse optical flow for moving objects	CV	No	SDK UI	SparseOpticalFlowAlgorithm
metavision_dense_optical_flow	Computes the optical flow for moving objects at each event along the edge’s normal	CV	No	SDK UI	PlaneFittingFlowAlgorithm TripletMatchingFlowAlgorithm
metavision_undistortion	Undistort and distort coordinates of events	CV	No	OpenCV	CameraGeometry
metavision_xyt	Displays events in a 3D space	CV	No	ImGui	TrailFilterAlgorithmT SpatioTemporalContrastAlgorithm AntiFlickerAlgorithm
metavision_active_marker_2d_tracking	Tracks Active Markers in 2D	CV	No	SDK UI	ModulatedLightDetectorAlgorithm ActiveMarkerTrackerAlgorithm
metavision_edgelet_2d_tracking	Tracks 2D edgelets	CV3D	No	OpenCV	Edgelet2dDetectionAlgorithm Edgelet2dTrackingAlgorithm
metavision_model_3d_tracking	Tracks known 3D objects	CV3D	No	SDK UI	Model3dDetectionAlgorithm Model3dTrackingAlgorithm
metavision_active_marker_3d_tracking	Tracks Active Markers in 3D	CV3D	No	Ogre	ModulatedLightDetectorAlgorithm ActiveMarkerPoseEstimatorAlgorithm
metavision_counting	Counts small objects moving vertically (e.g. bulk counting)	Analytics	No	SDK UI	CountingAlgorithm
metavision_jet_monitoring	Monitors jets being dispensed and sends alarms when the dispensing frequency isn’t correct	Analytics	No	SDK UI	JetMonitoringAlgorithm
metavision_psm	Counts and estimates the sizes of objects moving vertically	Analytics	No	SDK UI	PsmAlgorithm
metavision_generic_tracking	Tracks any moving object	Analytics	No	SDK UI	TrackingAlgorithm SpatioTemporalContrastAlgorithm
metavision_spatter_tracking	Tracks simple non-colliding objects	Analytics	No	SDK UI	SpatterTrackerAlgorithm SpatioTemporalContrastAlgorithm
metavision_vibration_estimation	Estimates vibration frequency per pixel and shows the dominant frequency (the most common frequency among all pixels)	Analytics	No	SDK UI	FrequencyMapAsyncAlgorithm
metavision_detection_and_tracking_pipeline	Inference Pipeline of Detection and Tracking	ML	No	OpenCV	DTPipeline

Advanced modules Python samples

Python samples of the advanced modules are showing some algorithms in action.

In the following table, for each sample, you will find its modules and the main algorithm or classes demonstrated in the sample.

Name	Description	Module	Main algorithms demonstrated in the sample
show_calibrated_poses	Visualize the 3D locations of a calibration pattern detected during intrinsics calibration	Calibration	N/A
metavision_ground_plane_calibration	Computes the mapping between the camera’s reference system and the world’s reference system (camera is rigidly attached)	Calibration	FrequencyAlgorithm FrequencyClusteringAlgorithm
metavision_noise_filtering	Filters events with a noise filter (Activity, Trail or STC) and displays them on the screen	CV	ActivityNoiseFilterAlgorithm TrailFilterAlgorithm SpatioTemporalContrastAlgorithm
metavision_data_rate	Displays event rate from an event-based camera or from a RAW file	CV	ActivityNoiseFilterAlgorithm SpatioTemporalContrastAlgorithm AntiFlickerAlgorithm
metavision_sparse_optical_flow	Computes the sparse optical flow for moving objects	CV	SparseOpticalFlowAlgorithm
metavision_dense_optical_flow	Computes the dense normal optical flow for moving objects at each event along the edge’s normal	CV	PlaneFittingFlowAlgorithm TripletMatchingAlgorithm
metavision_model_3d_tracking	Tracks known 3D objects	CV3D	Model3dDetectionAlgorithm Model3dTrackingAlgorithm
metavision_counting	Counts small objects moving vertically (e.g. bulk counting)	Analytics	CountingAlgorithm
metavision_jet_monitoring	Monitors jets being dispensed and sends alarms when the dispensing frequency isn’t correct	Analytics	JetMonitoringAlgorithm
metavision_psm	Counts and estimates the sizes of objects moving vertically	Analytics	PsmAlgorithm
metavision_generic_tracking	Tracks any moving object	Analytics	TrackingAlgorithm ActivityNoiseFilterAlgorithm TrailFilterAlgorithm
metavision_spatter_tracking	Tracks simple non-colliding objects	Analytics	SpatterTrackerAlgorithm SpatioTemporalContrastAlgorithm
metavision_vibration_estimation	Estimates vibration frequency per pixel and shows the dominant frequency (the most common frequency among all pixels)	Analytics	FrequencyMapAsyncAlgorithm
metavision_detection_and_tracking_pipeline	Inference Pipeline of Detection and Tracking	ML	DTPipeline DataAssociation
eval_coco_kpi	Evaluates Detection and Tracking pipeline on coco KPIs	ML	evaluate_detection
vizu_gt_det	Visualizes Ground Truth and ML detection results	ML	draw_box_events
train_detection	Trains Detection model with Pytorch Lightning	ML	LightningDetectionModel
export_detector	Exports detection checkpoint to torch.jit model	ML	LightningDetectionModel
classification_inference	Performs classification inference	ML	CDProcessorIterator HDF5Iterator
train_classification	Trains supervised classification model with Pytorch Lightning	ML	ClassificationModel
export_classifier	Exports classification checkpoint to torch.jit model	ML	ClassificationModel
flow_inference	Performs Optical Flow inference with a flow network	ML	FlowModel
train_flow	Trains Optical Flow model	ML	FlowModel
export_flow	Exports flow checkpoint to torch.jit model	ML	FlowModel
generate_hdf5	Writes HDF5 feature files from event files	ML	generate_hdf5
viz_data	Visualizes a precomputed HDF5 dataset with or without bbox labels	ML	SequentialDataLoader
viz_moving _mnist	Visualizes the Moving MNIST	ML	MovingMNISTDataset
bbox_txt2npy	Converts Bbox from text format to numpy format	ML	EventBbox

Samples listed by Module

• Analytics
• Calibration
• CV
• CV3D
• ML