object检测fps和视频fps有什么区别？

Question

`I'm using in edge impulse FOMO

I know that object detection fps is 1/inference time my model's time per inference is 2ms so object detection is 500fps

but my model run vscode fps is 9.5

what is the difference between object detection fps and video fps?

Answer 1

If I understand correctly, your object detection fps indicates the number of frames (or images) that your model, given your system, can process in a second.

A video fps in your input source's frames per second. For example, if your video has an fps (also referred to as framerate) of 100, then your model would be able to detect objects in all of the frames in 100ms (or 1/10 of a second).

In your case, your video input source seems to have 9.5 frames in a second. This means that your model, given your system, will process 1-second wort of a video in about ~20ms.