Micro-rigid registration

Classes

MicroRigidRegistrar

class valis.micro_rigid_registrar.MicroRigidRegistrar(val_obj, feature_detector_cls=<class 'valis.feature_detectors.SuperPointFD'>, matcher=<class 'valis.feature_matcher.SuperPointAndGlue'>, processor_dict=None, scale=0.125, tile_wh=512, roi='mask')[source]

Refine rigid registration using higher resolution images

Rigid transforms found during lower resolution images are applied to the WSI and then downsampled. The higher resolution registered images are then divided into tiles, which are processed and normalized. Next, features are detected and matched for each tile, the results of which are combined into a common keypoint list. These higher resolution keypoints are then used to estimate a new rigid transform. Replaces thumbnails in the rigid registration folder.

val_obj

The “parent” object that registers all of the slides.

Type:

Valis

feature_detector_cls

Uninstantiated FeatureDD object that detects and computes image features. Default is SuperPointFD. The available feature_detectors are found in the feature_detectors module. If a desired feature detector is not available, one can be created by subclassing feature_detectors.FeatureDD.

Type:

FeatureDD, optional

matcher

Matcher object that will be used to match image features

Type:

Matcher

scale

Degree of downsampling to use for the reigistration, based on the registered WSI shape (i.e. Slide.aligned_slide_shape_rc)

Type:

float

tile_wh

Width and height of tiles extracted from registered WSI

Type:

int

roi

Determines how the region of interest is defined. roi=”mask” will use the bounding box of non-rigid registration mask to define the search area. roi=matches will use the bounding box of the previously matched features to define the search area.

Type:

string

iter_order

Determines the order in which images are aligned. Goes from reference image to the edges of the stack.

Type:

list of tuples

__init__(val_obj, feature_detector_cls=<class 'valis.feature_detectors.SuperPointFD'>, matcher=<class 'valis.feature_matcher.SuperPointAndGlue'>, processor_dict=None, scale=0.125, tile_wh=512, roi='mask')[source]
Parameters:

  • val_obj (Valis) – The “parent” object that registers all of the slides.

  • feature_detector_cls (FeatureDD, optional) – Uninstantiated FeatureDD object that detects and computes image features. Default is SuperPointFD. The available feature_detectors are found in the feature_detectors module. If a desired feature detector is not available, one can be created by subclassing feature_detectors.FeatureDD.

  • matcher (Matcher) – Matcher object that will be used to match image features

  • processor_dict (dict, optional) – Each key should be the filename of the image, and the value either a subclassed preprocessing.ImageProcessor, or a list, where the 1st element is the processor, and the second element a dictionary of keyword arguments passed to the processor. If None, a default processor will be assigned to each image based on its modality.

  • scale (float) – Degree of downsampling to use for the reigistration, based on the registered WSI shape (i.e. Slide.aligned_slide_shape_rc)

  • tile_wh (int) – Width and height of tiles extracted from registered WSI

  • roi (string) – Determines how the region of interest is defined. roi=”mask” will use the bounding box of non-rigid registration mask to define the search area. roi=matches will use the bo

register(brightfield_processing_cls=<class 'valis.preprocessing.StainFlattener'>, brightfield_processing_kwargs={'adaptive_eq': False, 'with_mask': False}, if_processing_cls=<class 'valis.preprocessing.ChannelGetter'>, if_processing_kwargs={'adaptive_eq': True, 'channel': 'dapi'})[source]
Parameters:

  • brightfield_processing_cls (ImageProcesser) – ImageProcesser to pre-process brightfield images to make them look as similar as possible. Should return a single channel uint8 image.

  • brightfield_processing_kwargs (dict) – Dictionary of keyward arguments to be passed to brightfield_processing_cls

  • if_processing_cls (ImageProcesser) – ImageProcesser to pre-process immunofluorescent images to make them look as similar as possible. Should return a single channel uint8 image.

  • if_processing_kwargs (dict) – Dictionary of keyward arguments to be passed to if_processing_cls