Dear Participants,

Thanks to inquiries from pzcai, gepengai.ji, and frauke.wilm, I have realized there was some confusion regarding the COSAS Challenge. I would like to clarify the task details further.

The COSAS Challenge involves two-category image segmentation: negative and positive.

  • In the Ground Truth:
  • 0 represents the negative category.

  • Any value other than 0 represents the positive category.

  • In the Prediction:

  • 0 represents the negative category.
  • 1 represents the positive category.

Please note that for the Grand Challenge, the pixel values in the prediction image must be in the range from 0 to 3. If the pixel values fall outside this range, the platform will report an error.

I hope this clarification helps. If you have any further questions, please feel free to ask.

Best regards,

Xi Long