API Reference¶

`nondefaced_detector.models`: Model functions¶

`nondefaced_detector.models`
`nondefaced_detector.models.model.ConvBNrelu`(x)	A layer block of one convolutional, one batch normalization, and one non-linear activation sequence.
`nondefaced_detector.models.model.TruncatedSubmodel`(…)	The TruncatedSubmodel trained in Step 1 of the model.
`nondefaced_detector.models.model.ClassifierHead`(…)	The final block of the model
`nondefaced_detector.models.model.Submodel`(…)	3 identical submodel blocks are used to train on spatial information from all three axes (axial, coronal, sagittal) separately.
`nondefaced_detector.models.model.CombinedClassifier`([…])	The final block of the model that combines features and outputs a real-valued probability using the sigmoid function.

`nondefaced_detector.dataloaders`
`nondefaced_detector.dataloaders.dataset.get_dataset`(…)	Returns tf.data.Dataset after preprocessing from tfrecords for training and validation

Script to preprocess volumes

`nondefaced_detector.preprocess`	Script to preprocess volumes
`nondefaced_detector.preprocess.preprocess`(…)	Preprocess input volumes before prediction.
`nondefaced_detector.preprocess.preprocess_parallel`(…)	Preprocess multiple input volumes before prediction in parallel.

`nondefaced_detector.preprocessing`
`nondefaced_detector.preprocessing.conform.conform_data`(in_file)	Conform the input dataset to the canonical orientation.
`nondefaced_detector.preprocessing.normalization.clip`(x)
`nondefaced_detector.preprocessing.normalization.standardize`(x)
`nondefaced_detector.preprocessing.normalization.normalize`(x)

`nondefaced_detector.training`
`nondefaced_detector.training.training.train`(…)	Train a model.

Methods to predict using trained models

`nondefaced_detector.prediction`	Methods to predict using trained models
`nondefaced_detector.prediction.predict`(…)	Return predictions from a list of input volumes.
`nondefaced_detector.prediction._structural_slice`(x, …)	Transpose dataset and get slices from the volume based on the plane.
`nondefaced_detector.prediction._get_model`(…)	Return tf.keras.Model object from a filepath.

Standalone inference script for held-out test dataset.

`nondefaced_detector.inference`	Standalone inference script for held-out test dataset.
`nondefaced_detector.inference.inference`(…)	Inference function to reproduce original model scores.

`nondefaced_detector.helpers`
`nondefaced_detector.helpers.utils.is_gz_file`(…)
`nondefaced_detector.helpers.utils.save_vol`(…)	save_path: path to write the volume to tensor_3d: 3D volume which needs to be saved affine: image orientation, translation
`nondefaced_detector.helpers.utils.load_vol`(…)	load_path: volume path to load :returns: loaded 3D volume affine: affine data specific to the volume :rtype: volume
`nondefaced_detector.helpers.utils.imshow`(img1)

Utilities for Nondefaced-detector.

`nondefaced_detector.utils`	Utilities for Nondefaced-detector.
`nondefaced_detector.utils.get_datalad`([…])	Download a datalad dataset/repo.