Hi, and thanks for producing this dataset!
I'm wondering if you can point me to/release the code you used for processing the raw crawler files and produce the webui-boxes json files, that are used for training the object detector models? I guess there are some nitty gritty details on how you e.g. filter the boxes to include?