Skip to content

Conversation

@Nic-Ma
Copy link
Contributor

@Nic-Ma Nic-Ma commented Mar 5, 2020

Fixes #144 #146 .

Description

This PR implemented:

  1. Vanilla level and dictionary level transforms of LoadNifti.
  2. Common Dataset for dictionary format input data.
  3. Updated 3D segmentation example to the latest dictionary-based APIs.
  4. Added ChangeToChannelFirst transform.
  5. Fixed bug in Dice loss.

Status

Ready

Types of changes

  • Bug fix (non-breaking change which fixes an issue)
  • Breaking change (fix or new feature that would cause existing functionality to change)
  • New tests added to cover the changes
  • Docstrings/Documentation updated

@Nic-Ma
Copy link
Contributor Author

Nic-Ma commented Mar 5, 2020

Hi @wyli @ericspod @atbenmurray ,

I updated code according to the comments.
Could you please help review it again?
Thanks.

Copy link
Contributor

@madil90 madil90 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@Nic-Ma Left comments inline.

@Nic-Ma
Copy link
Contributor Author

Nic-Ma commented Mar 6, 2020

Hi @wyli @madil90 @ericspod @atbenmurray ,

I updated the code according to the comments.
And also added below features to this PR:

  1. Added ChangeToChannelFirst transform.
  2. Updated 3D segmentation example to the latest dictionary-based APIs.
    Used ChangeToChannelFirstd, RandCropByPosNegLabeld, RandRotate90d transforms.
  3. Fixed bug in Dice loss: remove smooth from top, otherwise it may generate negative loss value.

I think we need another separate PR to update inference related logic.
Could you please help review it again?
Thanks.

@Nic-Ma Nic-Ma linked an issue Mar 6, 2020 that may be closed by this pull request
@madil90 madil90 merged commit 39b2cb2 into master Mar 9, 2020
@wyli wyli deleted the 144-loadnifti-transform-common-dataset branch May 21, 2020 13:31
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

update train example with transforms loadnifti transform and common dataset

4 participants