search

yijingru / BBAVectors-Oriented-Object-Detection

[WACV2021] Oriented Object Detection in Aerial Images with Box Boundary-Aware Vectors
MIT License
469 stars 88 forks source link

Train on custom data #100

Open meiqing-wang opened 3 years ago

meiqing-wang commented 3 years ago

Hi~ I tested your codes on DOTA dataset and the results looked good. So I am gonna apply your method to my dataset. When preparing the labels for my data, I found there were class category and difficulty in labels. In my case I only have one category and I don't know how to label 'difficulty' as it seems there is no labelling tool for this parameter. Can I just remove difficulty and set the number of class as 1? Will these settings impact the results a lot?

Looking forward to your reply! :)

yijingru commented 3 years ago

Hi, I think I didn't really use 'difficulty'. You may remove the key or set it to any number.