Closed Xanthorapedia closed 4 years ago
Thanks for liking it. As for the dataset, I will consider it. The main concern is that that dataset is basically a bunch of photos of me shooting basketball, which means a lot of my stupid faces are included. That's kind of a big problem...
I see. Very reasonable concern. I personally wouldn't do that either. However, can you describe your workflow for collecting the dataset?
It's actually really simple. I used my phone to took about 500 photos on the court, and manually labeled them all.
It's actually really simple. I used my phone to took about 500 photos on the court, and manually labeled them all. Could you give me an example of your data, just a few pieces and how to make labels ?Photos can be blurred
I wonder if you will be publishing your dataset for detection. Thanks. Impressive work by the way!