0
$\begingroup$

I want to use darkflow to perform object detection. I have 5 classes of objects I want to detect, say ,

  chair
  bottle
  door
  phone
  computer

I want to specifically train my model again, not use pre-trained weights. I plan to train it on images which are annotated with bounding boxes. But all available datasets like imagenet, COCO have images that have objects of other classes also annotated in them. So how can I use these datasets to train specifically on my 5 classes? I don't want my model training on all the other objects in the images.

$\endgroup$

2 Answers 2

1
$\begingroup$

You need to change xml or csv files of bounding boxes and from there you need to remove the all classes other then your 5 classes.

$\endgroup$
0
$\begingroup$

Take one of those data sets, and remove all labels/annotations for any class other than the 5 you care about.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.