YoloV9 overpredicting tons of false positives

Hi there,

I’m pretty new to the topic of ML at all, currently just learning all the “boring” mathematic and statistical basics. But as a side project I wanted to get in touch with the new Yolo and do a detection of our pets.

So, at the first starts, I labeled like 20 pictures of them and did some prediction with videos of them, and it recognized in some rare, but not all cases the rabbits. Nevertheless, somehow the model always thought that the food bowl is one of the pets :slight_smile:

So, I thought now to tag more example images of them, having now like around 50 instead of 20 with my first test. But now I have many, many false positives no matter which video I test with…

What am I essentially doing wrong here?

The current dataset can be found here:

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.