NVIDIA's Jetson is a line of cutting-edge GPUs that allow roboticists to work with video and image data in ways that were once practically unimaginable. Combined with its machine-learning powered Deepstream SDK, Jetson gives you a veritable treasure trove of capabilities that make working with visual robotic data a finely-tuned art. Formant leverages both technologies to deliver infinitely more accessible, observable, and analyzable robot data.
OPTIMIZE PROCESSING WITH
By virtue of its novel GPU and hardware-accelerated plugins, Jetson encodes video at the edge — this frees up CPU for other tasks. Formant also adds H.264 encoding, which provides high-quality transmission of full-motion video, lower bandwidth requirements, lower latency, and lower storage requirements.
AUTOMATE PRIVACY WITH DE-IDENTIFICATION AT THE EDGE
With privacy concerns on the rise, and governments worldwide scrutinizing the use of facial-recognition software, it’s imperative that one must take ethical and legal compliance seriously when working with visual data streaming from robots. Formant leverages NVIDIA de-identification technology to anonymize video streams on-robot, so faces are obscured BEFORE they are ingested to the cloud.
Enrich insights with semantic identification
Formant uses NVIDIA Deep Stream to make your data smarter. Semantic identification enables objects captured by cameras to be automatically identified and labelled. This creates an immense, and ever-expanding, library of visual objects which can be used in downstream analytics, machine-learning or other workflows/applications.