US 2022/0067944 A1 · SHIFT INVARIANT LOSS FOR DEEP LEARNING BASED IMAGE SEGMENTATION

Abstract

Systems and methods of improving alignment in dense prediction neural networks are disclosed. A method includes identifying, at a computing system, an input data set and a label data set with one or more first parts of the input data set corresponding to a label. The computing system processes the input data set using a neural network to generate a predicted label data set that identifies one or more second parts of the input data set predicted to correspond to the label. The computing system determines an alignment result using the predicted label data set and the label data set and a transformation of the one or more first parts, including a shift, rotation, scaling, and/or deformation, based on the alignment result. The computing system computes a loss score using the transformation, label data and the predicted label data set and updates the neural network based on the loss score.

Claims21

Dependent claims

A method comprising:

receiving a first image of an unstained tissue sample;
receiving a second image of the unstained tissue sample;
generating, by a trained neural network, a first set of labels corresponding to first tissue features within the first image, and a second set of labels corresponding to second tissue features within the second image;
determining, by the trained neural network, an alignment result between the first and second images based at least in part on at least one of the first and second sets of labels;
generating, by the trained neural network, an aligned first image on the alignment result;
and
applying a virtual stain to the aligned first image.

4 more elements

A system comprising:

a non-transitory computer-readable medium;
and
one or more processors communicatively coupled to the non-transitory computer-readable medium, the one or more processors configured to execute processor-executable instructions stored in the non-transitory computer-readable medium to:
receive a first image of an unstained tissue sample;
receive a second image of the unstained tissue sample;
generate, by a trained neural network, a first set of labels corresponding to first tissue features within the first image, and a second set of labels corresponding to second tissue features within the second image;
determine, by the trained neural network, an alignment result between the first and second images based at least in part on at least one of the first and second sets of labels;
generate, by the trained neural network, an aligned first image on the alignment result;
and
apply a virtual stain to the aligned first image.

7 more elements

Dependencies are inferred from the claim text.

Shift invariant loss for deep learning based image segmentation

Life of this patent

Shift invariant loss for deep learning based image segmentation

Life of this patent

Abstract

Claims21