So as far as I understand, when you have a trained YOLO model, the final prediction looks like an image with couple bounding boxes with class names objects belong to.
The question is: is it possible to save these names of classes that were detected on image(for example as a list) for future analysis?
Anything would be helpful, I don't care if it can be done in python, .net or whatever.
I'm completely new to object detection, would really appreciate your help :)
Yes, it is possible and thats exactly what you receive. To clarify further, what you see as the output as a saved image with bounding boxes is what most scripts generate so they can show you the end results. In reality however, the output of every trained detector is to give you the object classes and the associated points for the bounding box (these bounding boxes are given out in different formats, but thats a huge discussion so best you google that directly). These bounding boxes and class labels are then drawn over the image and then that whole image is saved with this data. The obtained classes and bounding boxes can just be written to a python dataframe which can then be exported to a CSV.
TL;DR - The output of all networks is actually the classes and the bounding boxes (not the image) which can then directly be stored as a python dataframe and converted to a CSV.
