SynAM-E
The first publicly available event-camera benchmark dataset in the metal additive manufacturing domain. It comprises 85 physically calibrated event-data segments drawn from 15 data sources across 8 institutions, and is designed to evaluate cross-device generalization of event cameras on melt-pool monitoring tasks.
Timeline (last 90 days)
SynAM-E is the first publicly available event camera benchmark dataset in the field of melt pool monitoring for metal additive manufacturing.
SynAM-E contains 85 physically calibrated event data segments from 15 data sources across 8 institutions.
SynAM-E provides publicly available baseline models and fixed cross-machine evaluation splits.
In a single-machine case study, the macro F1 metric for event-based spatial monitoring reached 0.874, roughly on par with the dense frame solution's 0.863.
In the trust test, process-type classification passed, but material classification remained confounded by camera spectral band.
All Facts (5)
SynAM-E is the first publicly available event camera benchmark dataset in the field of melt pool monitoring for metal additive manufacturing.
50%UnverifiedSynAM-E contains 85 physically calibrated event data segments from 15 data sources across 8 institutions.
50%UnverifiedSynAM-E provides publicly available baseline models and fixed cross-machine evaluation splits.
50%UnverifiedIn a single-machine case study, the macro F1 metric for event-based spatial monitoring reached 0.874, roughly on par with the dense frame solution's 0.863.
50%UnverifiedIn the trust test, process-type classification passed, but material classification remained confounded by camera spectral band.
50%