3 ms·
FTR, a simple RPi4 yields ~4FPS on a DL object detection model. You may need cloud for your central hub, but neural network inference can be done locally.
by pilooch 7y ago
FTR, a simple RPi4 yields ~4FPS on a DL object detection model. You may need cloud for your central hub, but neural network inference can be done locally.
- mrafiee 7y agoBy 4FPS do you mean 250ms latency per frame? What OD model specifically did you get that latency from on RPi4?