Software·43,989 discussions

Deploying Object Tracking Models on Raspberry Pi Using TensorFlow Lite

Tags: Software

Trying to deploy a multi-object tracking model using TensorFlow Lite on a Raspberry Pi 4. Having trouble with frame rates and accuracy. Has anyone done this efficiently with low latency?

Topic

Reply To: Deploying Object Tracking Models on Raspberry Pi Using TensorFlow Lite

In reply to Deploying Object Tracking Models on Raspberry Pi Using TensorFlow Lite

https://github.com/EdjeElectronics/TensorFlow-Lite-Object-Detection-on-Android-and-Raspberry-Pi/blob/master/deploy_guides/Raspberry_Pi_Guide.md , alas, doesn’t say anything about frames rate and accuracy.

But it could be you’re reaching the limits of your hardware.

From what I’ve experienced.

In reply to Deploying Object Tracking Models on Raspberry Pi Using TensorFlow Lite

This is a low resource platform so expect low rates and along with that some accuracy problems.

Nothing broken here. Try again on a platform suitable for the work.

[Author]

Replies