Real time body segmentation on coral

TensorFlow General Discussion
,
1

Hi experts
I’m currently using Jetson nano 4gb for a real time body Jetson segmentation model. For the moment the best I seem to get is around 4 fps on a 1280*720 image with approx 80% accuracy.
I would like to know if I get better performance with coral? I’m aiming towards 30 fps if possible.
I don’t have any preference in terms of a specific model. All I really care about is the image resolution, fps and a decent accuracy.

Thanks!

3

Hi @Tsino ,

Coral is likely a better choice for real-time body segmentation due to its superior performance and suitability for high-resolution images and high FPS requirements. Consider factors like model complexity, image resolution, FPS, accuracy, cost, power consumption, and development environment when making your decision.

Thank You .