Recordings and Datasets

Sample recordings

Those recordings are available in multiple formats: RAW (EVT2 or EVT3), DAT and HDF5 event files.

Name	Description	Sensor	Duration	Recordings
spinner	Turning Spinner	Gen3.0	5 s	RAW EVT2 (207.9 MB) HDF5 (152.8 MB) DAT (413.2)
hand_spinner	Turning 3-Blade Hand Spinner	Gen3.1	5 s	RAW EVT2 (46.9 MB) HDF5 (38.7 MB)
laser	High-speed laser motion	Gen4.1	3 s	RAW EVT3 (326.2 MB) HDF5 (236.5 MB)
80_balls	Falling 80 balls	Gen3.0	6.3 s	RAW EVT2 (19.2 MB) HDF5 (14.8 MB)
195_falling_particles	Falling 195 particles of different sizes (synthetic data)	Gen3.1	32 ms	RAW EVT2 (77.3 MB) HDF5 (58.2 MB)
monitoring_40_50hz	Object vibration at 40 and 50Hz	Gen3.0	6 s	RAW EVT2 (156.8 MB) HDF5 (112.8 MB)
sparklers	High-speed scatter-like motion of small particles	Gen3.0	95 ms	RAW EVT2 (2 MB) HDF5 (1.5 MB)
200_jets_at_200hz	200 horizontal fluid jets	Gen3.1	2 s	RAW EVT2 (2.3 MB) HDF5 (1.5 MB)
pedestrians	Pedestrians in multiple directions	Gen4.1	30 s	RAW EVT3 (183.7 MB) HDF5 (126.6 MB)
traffic_monitoring	Cars moving on a highway (static camera)	Gen3.0	28 s	RAW EVT2 (70.3 MB) HDF5 (62.5 MB)
driving_sample	Driving on a street road (moving camera)	Gen4.1	12.5 s	RAW EVT3 (352.2 MB) HDF5 (270.2 MB)
box	Camera moving around a box	Gen3.1	28.6 s	ZIP EVT2 (645.1 MB)
cube	Camera moving around a cube	Gen3.1	17.7 s	ZIP EVT2 (378.3 MB)
marker	Camera moving around a marker	Gen3.1	22.3 s	ZIP EVT2 (339.2 MB)

Dataset	Application	Sensor	Resolution	Size	Format
1 Megapixel Automotive Detection	Automotive detection	Gen4.0	1280×720	1.6T compressed 3.5T uncompressed	Recording in DAT format Labels in numpy format
HVGA ATIS Corner	Corner detection	Gen3.0 ATIS	480×360	6G compressed 11G uncompressed	Recording in DAT format
GEN1 N-CARS	Car classification	Gen1.0	crops from 304×240	280M compressed	Recording in DAT format
GEN1 Automotive Detection	Automotive detection	Gen1.0	304×240	200G compressed 750G uncompressed	Recording in DAT format Labels in numpy format
Simplified mini-dataset	ML Quick Start Tutorial	Gen4.0	1280×720	24G uncompressed	Recording in DAT format Labels in numpy format

Those datasets were precompiled in an HDF5 tensor file format for faster training and smaller disk usage.

Dataset	Preprocessing	Sensor	Resolution	Size	Format
Gen 4 automotive dataset	Event Cube	Gen4.0	1280×720 down to 640x360	211.3G	HDF5 Labels in numpy format
Gen 4 automotive dataset	Multi Channel Timesurface	Gen4.0	1280×720 down to 640x360	288.4G	HDF5 Labels in numpy format
Gen 4 automotive dataset	Histograms	Gen4.0	1280×720 down to 640x360	86G	HDF5 Labels in numpy format
Gen 1 automotive dataset	Event Cube	Gen1.0	304×240	59.3GB	HDF5 Labels in numpy format