DeepLIIF 1.2.5

• Fallback to black pixels if failure in reading WSI tile

DeepLIIF 1.2.4

• Revised DeepLIIF model class, allowing for a variable number of modalities to be used for training and inference.

DeepLIIF 1.2.3

V1.2.3

Bump setup version number

DeepLIIF 1.2.2

• Draw outlines outside of segmented cells instead of on boundary.
• Fallback to BioFormats if TiffFile fails in reading WSI tile.

DeepLIIF 1.2.1

• Do not threshold out large noise regions by default.

DeepLIIF 1.2.0

• Update default segmentation weights

DeepLIIF 1.1.15

• Add seg_weights parameter for inference functions

DeepLIIF 1.1.14

• Add knowledge distillation model class
• Add large noise threshold for postprocessing

DeepLIIF 1.1.13

• Remove OpenSlide dependencies
• Upgrade default PyTorch version

DeepLIIF 1.1.12

• Add --seg-only option to create only images necessary for final segmentation, reducing overall run time

• Add --seg-intermediate option to output intermediate individual modality segmentation maps

• Add CycleGAN model class

• Add postprocessing function for individual cell output