how to run

[Eugenel00]

Hi, great work, can you share the preprocessing dataset used, it would be better if you could explain the steps to train the model, looking forward to your reply

[CaelynCheung1996]

Hi, great work, can you share the preprocessing dataset used, it would be better if you could explain the steps to train the model, looking forward to your reply

Hi Eugenel,

Thanks for you interest in our work. Unfortunately, we are not able to release the preprocessing dataset due to confidential issue. If you have exact similar data, I suggest you to try our pretrained model and see how it performs. Otherwise, you can refer to the matterport documentation for more training details to train from scratch. The postprocessing code on our repository should work perfectly on the output of Mask RCNN if you are also working with imaging of CLS structure.

[Eugenel00]

Thank you very much for your reply and suggestion, could you advise the data file format of the input of the preprocessing code if possible? Where “root_dir_masks = '/path/to/label/'；root_dir_images = '/path/to/image'”.
@CaelynCheung1996

[CaelynCheung1996]

Thank you very much for your reply and suggestion, could you advise the data file format of the input of the preprocessing code if possible? Where “root_dir_masks = '/path/to/label/'；root_dir_images = '/path/to/image'”. @CaelynCheung1996

For the light-sheet images, the format is ".tif" and the corresponding mask format is ".png."