zhangye-zoe / BASS

5 stars 0 forks source link

dataset #1

Open zaoruis opened 1 day ago

zaoruis commented 1 day ago

image Thank you for your work. What do these filenames stand for? My data set is in coco format, what do I need to convert it?

zhangye-zoe commented 1 day ago

We have uploaded the data processing code and file to the google drive: https://drive.google.com/drive/folders/1u5LRyaF5qdWZ7No3NPsqvZVuxz46dRhn?usp=drive_link

  1. In the directory, you can see the original dataset download from the official website is saved in Origin;
  2. Then, you can run "data.ipynb" file to split each image 256*256 patches, which will be saved in train/val/test directory;
  3. Last, you can run "generate_coco_format.ipynb" file to generate the training/validation/test datasets.