image_classifier_3d package

Subpackages

• image_classifier_3d.bin package
  • Submodules
  • image_classifier_3d.bin.run_classifier module
  • Module contents
• image_classifier_3d.data_loader package
  • Submodules
  • image_classifier_3d.data_loader.universal_loader module
  • image_classifier_3d.data_loader.utils module
  • Module contents
• image_classifier_3d.models package
  • Submodules
  • image_classifier_3d.models.build_classifier module
  • image_classifier_3d.models.densenet module
  • image_classifier_3d.models.resnet module
  • image_classifier_3d.models.resnet_GN module
  • Module contents
• image_classifier_3d.utils package
  • Submodules
  • image_classifier_3d.utils.misc_utils module
  • image_classifier_3d.utils.quilt_utils module
  • Module contents

Submodules

image_classifier_3d.proj_tester module

class image_classifier_3d.proj_tester.ProjectTester(mode: str = 'inference', save_model_output: bool = False)

Bases: object

Main class for applying a new classifier on evaluation data or new data

initialization

mode: str

running mode, either “inference” (no ground truth) or “evaluation” (ground truth is available)

save_model_output: bool

whether to save the raw prediction from model(s)

run_tester_config(config_filename: Union[str, pathlib.Path], output_path: Union[str, pathlib.Path], return_df: bool = False)

do testing using a config yaml file

config_filename: Union[str, Path]

the configuration file to use

output_path: Union[str, Path]

where to save the prediction outputs (as a CSV file)

return_df: bool

whether a dataframe will be returned besides saving to CSV, default is False

run_tester_csv(csv_filename: Union[str, pathlib.Path], output_path: Union[str, pathlib.Path], return_df: bool = False, project_name: str = 'mitotic_classifier', config_yaml: Union[str, pathlib.Path] = 'default') → Optional[pandas.core.frame.DataFrame]

do testing using data from a csv file

csv_filename: Union[str, Path]

the filepath to the csv file

output_path: Union[str, Path]

where to save the prediction outputs (as a CSV file)

return_df: bool

whether a dataframe will be returned besides saving to CSV, default is False

project_name: str

which project (i.e., which kind of classifier) to use

config_yaml: Union[str, Path]

the configuration file to use, if it is “default”, the a yaml at ../model_zoo/test_config_{project_name}.yaml is used by default

image_classifier_3d.proj_trainer module

class image_classifier_3d.proj_trainer.ProjectTrainer(config_filename: Union[str, pathlib.Path])

Bases: object

Main class for training a new classifier project

Parameters

config_filename (Union(str,Path)) – path to the configuration file (.yaml)

run_trainer()

do the training

Module contents

Top-level package for Image Classifier 3D.

image_classifier_3d.get_module_version()