Validation of my Object Detection results and action required?

I am building a vehicle detection system and I have loaded around 1000 images and categorised them.

Looking at the “Object Detection” results, am I understanding them correctly, in that I need to upload a background image collection to make the detection more accurate as I haven’t included specific background images in my data set.

Please see attached image.

Hi @c.cooper

Hmmm yes it looks like the background class is under represented here, see our guide here for more steps - Increasing model performance | Edge Impulse Documentation

Best

Eoin