APPARATUS AND A METHOD FOR ON-DEMAND REGISTRATION OF WHOLE SLIDE IMAGES

1 . An apparatus for on-demand registration of whole slide images (WSIs), wherein the apparatus comprises:

at least a processor; and

a memory communicatively connected to the at least a processor, wherein the memory contains instructions configuring the at least a processor to:

receive at least two digital images, each comprising a pyramid structure, wherein the pyramid structure comprises:

a plurality of layers representing same image content at different magnification levels; and

a first layer having an image with a first resolution and a second layer having an image with a second resolution;

identify a first region of interest (ROI) on a first digital image of the at least two digital images at a target magnification;

determine a transformation matrix as a function of at least a portion of the first digital slide image and at least a portion of a second digital image of the at least two digital images at a first magnification level corresponding to another one of the plurality of layers of the pyramid structure;

identify a second ROI on the second digital image based on applying the transformation matrix to the first ROI; and

register the second ROI to the first ROI by applying the transformation matrix at the target magnification level.

2 . The apparatus of claim 1 , wherein the first and second digital images comprise whole slide images obtained from serially sectioned slides prepared from a common biological specimen.

3 . The apparatus of claim 1 , wherein identifying the first region of interest (ROI) comprises:

generating a user interface displaying the first digital image at the target magnification level; and

receiving a selection of the first ROI from the user interface.

4 . The apparatus of claim 1 , wherein determining the transformation matrix comprises:

identifying key points within the first ROI on the first digital image and the second ROI on the second digital image;

extracting descriptors from the key points that are invariant to scale;

matching the key points based on distances between the descriptors; and

computing a transformation based on their spatial correspondence.

5 . The apparatus of claim 1 , wherein the transformation matrix comprises a plurality of transformation parameters, wherein the plurality of transformation parameters comprises a scaling parameter for adjusting a size of an image.

6 . The apparatus of claim 1 , wherein the first magnification level is higher than the target magnification level.

7 . The apparatus of claim 1 , wherein registering the second ROI to the first ROI comprises:

receiving the first and second digital images as inputs to an optical flow model;

predicting a movement of pixels from the first digital image to the second digital image using the optical flow model; and

applying the movement of pixels to the second digital image.

8 . The apparatus of claim 1 , wherein registering the second ROI to the first ROI comprises:

detecting a failure of the registration of the second ROI to the first ROI at the target magnification level;

identifying a lower magnification level at which the registration is successful as a function of the failure;

performing a bottom-up registration process from the lower magnification level; and

deriving the second ROI on the second digital image as a function of the lower magnification level.

9 . The apparatus of claim 1 , wherein registering the second ROI to the first ROI comprises generating a visual alignment of the first and second ROIs on a digital pathology viewer at the target magnification level.

10 . The apparatus of claim 1 , wherein applying the transformation matrix comprises:

retrieving coordinate points defining a boundary of the first ROI;

performing matrix multiplication between the transformation matrix and each coordinate point; and

generating the second ROI on the second digital image using the resulting transformed coordinate points.

11 . A method for on-demand registration of whole slide images (WSIs), wherein the method comprises:

receiving, using at least a processor, at least two digital images, each comprising a pyramid structure, wherein the pyramid structure comprises:

a plurality of layers representing same image content at different magnification levels; and

a first layer having an image with a first resolution and a second layer having an image with a second resolution;

identifying, using the at least a processor, a first region of interest (ROI) on a first digital image of the at least two digital images at a target magnification;

determining, using the at least a processor, a transformation matrix as a function of at least a portion of the first digital slide image and at least a portion of a second digital image of the at least two digital images at a first magnification level corresponding to another one of the plurality of layers of the pyramid structure;

identifying, using the at least a processor, a second ROI on the second digital image based on applying the transformation matrix to the first ROI; and

registering, using the at least a processor, the second ROI to the first ROI by applying the transformation matrix at the target magnification level.

12 . The method of claim 11 , wherein the first and second digital images comprise whole slide images obtained from serially sectioned slides prepared from a common biological specimen.

13 . The method of claim 11 , wherein identifying the first region of interest (ROI) comprises:

generating a user interface displaying the first digital image at the target magnification level; and

receiving a selection of the first ROI from the user interface.

14 . The method of claim 11 , wherein determining the transformation matrix comprises:

identifying key points within the first ROI on the first digital image and the second ROI on the second digital image;

extracting descriptors from the key points that are invariant to scale;

matching the key points based on distances between the descriptors; and

computing a transformation based on their spatial correspondence.

15 . The method of claim 11 , wherein the transformation matrix comprises a plurality of transformation parameters, wherein the plurality of transformation parameters comprises a scaling parameter for adjusting a size of an image.

16 . The method of claim 11 , wherein the first magnification level is higher than the target magnification level.

17 . The method of claim 11 , wherein registering the second ROI to the first ROI comprises:

receiving the first and second digital images as inputs to an optical flow model;

predicting a movement of pixels from the first digital image to the second digital image using the optical flow model; and

applying the movement of pixels to the second digital image.

18 . The method of claim 11 , wherein registering the second ROI to the first ROI comprises:

detecting a failure of the registration of the second ROI to the first ROI at the target magnification level;

identifying a lower magnification level at which the registration is successful as a function of the failure;

performing a bottom-up registration process from the lower magnification level; and

deriving the second ROI on the second digital image as a function of the lower magnification level.

19 . The method of claim 11 , wherein registering the second ROI to the first ROI comprises generating a visual alignment of the first and second ROIs on a digital pathology viewer at the target magnification level.

20 . The method of claim 11 , wherein applying the transformation matrix comprises:

retrieving coordinate points defining a boundary of the first ROI;

performing matrix multiplication between the transformation matrix and each coordinate point; and

generating the second ROI on the second digital image using the resulting transformed coordinate points.