Ventana Medical Systems, Inc.

Fujifilm Corporation

Hoffmann-la Roche Inc.

Genentech, Inc.

Yao Nie

Maria V Sainz De Cea

Srinivas Chukka

Jim F Martin

Safoora Yousefi

Michael S Barnes

Ting Chen

Joerg Bredno

William Kenneth Day

Christophe Chefd'hotel

Daniel Bauer

Noemi Sebastiao

Akira Hasegawa

Eldad Klaiman

Xiangxue Wang

Fabien Gaire

Chao Shi

Auranuch Lorsakul

Alisa Tubbs

Jacob Gildenblat

Christian Roessler

Ido Ben-shaul

Xingwei Wang

Stanley Ho

Ofir Etz Hadar

Nariman Majdi-Nasab

Qinle Ba

Mohammadhassan Izady Yazdanabadi

Satarupa Mukherjee

Zuo Zhao

Xiao Li

Kien Nguyen

Nazim Shaikh

Trung Kien Nguyen

Robert Ochs

Alireza Chaman Zar

Michael Barnes

The present disclosure relates to automated systems and methods adapted to quickly and accurately train a neural network to detect and/or classify cells and/or nuclei. The present disclosure also relates to automated systems and methods for using a trained cell detection and classification engine, such as one including a neural network, to classify cells within an unlabeled image.

Systems and methods for identifying cell clusters within images of stained biological samples

A method and system for classifying field of view (FOV) images of histological slides into various categories that include certain stain patterns, artifacts, and/or other features of interest are provided herein. Few-shot learning (e.g., a prototypical network) techniques are used to train a deep convolutional neural network using a small number of training samples for a small number of image classes for classifying stain images belonging to a larger number of image classes.

Histological stain pattern and artifacts classification using few-shot learning

Systems and methods relate to predicting disease progression by processing digital pathology images using neural networks. A digital pathology image that depicts a specimen stained with one or more stains is accessed. The specimen may have been collected from a subject. A set of patches are defined for the digital pathology image. Each patch of the set of patches depicts a portion of the digital pathology image. For each patch of the set of patches and using an attention-score neural network, an attention score is generated. The attention-score neural network may have been trained using a loss function that penalized attention-score variability across patches in training digital pathology images labeled to indicate no or low subsequent disease progression. Using a result-prediction neural network and the attention scores, a result is generated that represents a prediction of whether or an extent to which a disease of the subject will progress.

Attention-based multiple instance learning

Methods and systems can include: accessing a digital pathology image; generating, using a first machine-learning model, a segmented image that identifies at least: a predicted diseased region and a background region in the digital pathology image; detecting depictions of a set of cells in the digital pathology image; generating, using a second machine-learning model, a cell classification for each cell of the set of cells, wherein the cell classification is selected from a set of potential classifications that indicate which, if any, of a set of biomarkers are expressed in the cell; detecting that a subset of the set of cells are within the background region; and updating the cell classification for each cell of at least some cells in the subset to be a background classification that was not included in the set of potential classifications.

Hybrid and accelerated ground-truth generation for duplex arrays

Aspects of the present disclosure pertain to systems and methods for enhancing brightfield or darkfield images to better enable nucleus detection. In some embodiments, the systems and methods described herein are useful for identifying membrane stain biomarkers as well as nuclear/cytoplasm stain biomarkers in stained images of biological samples. In some embodiments, the presently disclosed systems and methods enable quick and accurate nucleus detection in stained images of biological samples, especially for original stained images of biological samples where the nuclei appear faint.

Image enhancement to enable improved nuclei detection and segmentation

The present disclosure relates to techniques for transforming digital pathology images obtained by different slide scanners into a common format for image analysis. Particularly, aspects of the present disclosure are directed to obtaining a source image of a biological specimen, the source image is generated from a first type of scanner, inputting into a generator model a randomly generated noise vector and a latent feature vector from the source image as input data, generating, by the generator model, a new image based on the input data, inputting into a discriminator model the new image, generating, by the discriminator model, a probability for the new image being authentic or fake, determining whether the new image is authentic or fake based on the generated probability, and outputting the new image when the image is authentic.

Correcting differences in multi-scanners for digital pathology images using deep learning

Disclosed herein are systems and methods for of assessing stain titer levels. An exemplary method includes generating a set of field of views for the image or the region of the image, selecting field of views from the set of field of views that meet predefined criteria, creating a series of patches within each of the selected field of views, retaining patches from the series of patches that meet predefined criteria indicative of a presence of the stain for which the titer is to be estimated, deriving stain color features and stain intensity features pertaining to the stain from the retained patches, estimating a titer score for each of the retained patches based on the stain color features and the stain intensity features, and calculating a weighted average score for the titer of the stain based on the estimated titer score for each of the retained patches.

Automatic assay assessment and normalization for image processing

The present disclosure relates to techniques for segmenting and detecting cells within image data using transfer learning and a multi-task scheduler. Particularly, aspects of the present disclosure are directed to accessing a plurality of images of one or more cells, extracting three labels from the plurality of images, where the three labels are extracted using a Voronoi transformation, a local clustering, and application of repel code, training, by a multi-task scheduler, a convolutional neural network model based on three loss functions corresponding to the three labels, generating, by the convolutional neural network model, a nuclei probability map and a background probability map for each of the plurality of images based on the training with the three loss functions, and providing the nuclei probability map and the background probability map.

Weakly supervised multi-task learning for cell detection and segmentation

Disclosed herein are systems and methods for normalizing the titer of a first stain to a titer of the same stain in a template image. Also disclosed are methods of assessing stain titer levels.

An image analysis system for analyzing biological specimen images is disclosed. The system may include: a superpixel generator configured to obtain a biological specimen image and group pixels of the biological specimen image into a plurality of superpixels; a feature extractor configured to extract, from each superpixel in the plurality of superpixels, a feature vector comprising a plurality of image features; a clustering engine configured to assign the plurality of superpixels to a predefined number of clusters, each cluster being characterized by a centroid vector of feature vectors of superpixels assigned to the cluster; and a storage interface configured to store, for each superpixel in the plurality of superpixels, clustering information identifying the one cluster to which the superpixel is assigned. The system may also include a graph engine configured construct a graph based on the stored information, and use the graph to perform a graph-based image processing task.

Systems and methods for encoding image features of high-resolution digital images of biological specimens

The present disclosure provides systems and methods for separating colors in an image by automatically selecting color reference vectors that take into consideration the effect of light scattering, and principally how the light scattering changes the proportions of RGB channel signals in detected light at varying stain concentrations.

Color unmixing with scatter correction

Growing community of inventors

Sunnyvale, CA, United States of America

Yao Nie