Siemens Healthcare Diagnostics Gmbh

Covera Health, Inc.

JinHyeong Park

Terrence Chen

Benjamin S Pollack

Benjamin L Odry

Yao-Jen Chang

Wen Wu

Daniel Robert Elgort

Ron Vianu

Richard J Herzog

James Robert Browning

Benjamin Sellman Suutari

Bryce Eakin

Gregory Allen Dubbin

Tarmo Henrik Aijo

Xiaojin Dong

For training data pairs comprising training text (a radiological report) and training images (radiological images associated with the radiological report), a first encoder network determines word embeddings for the training text. A concept is generated from the operation of layers of the first encoder network, which is regularized by a first loss between the generated concept and a labeled concept for the training text. A second encoder network determines features for the training image. A heatmap is generated from the operation of layers of the second encoder network, which is regularized by a second loss between the generated heatmap and a labeled heatmap for the training image. A categorical cross entropy loss is calculated between a diagnostic quality category (classified by an error encoder) and a labeled diagnostic quality category for the training data pair. A total loss function comprising the first, second, and categorical cross entropy losses is minimized.

Computer-implemented machine learning for detection and statistical analysis of errors by healthcare providers

A model-based method of inspecting a specimen for presence of one or more interferent, such as Hemolysis, Icterus, and/or Lipemia (HI L) is provided. The method includes generating a pixelated image of the specimen in a first color space, determining color components (e.g., an a-value and a b-value) for pixels in the pixelated image, classifying of the pixels as being either liquid or non-liquid, defining one or more liquid regions based upon the pixels classified as liquid, and determining a presence of one or more interferent within the one or more liquid regions. The liquid classification is based upon a liquid classification model. Pixel classification may be based on a trained multiclass classifier. Interference level for the one or more interferent may be provided. Testing apparatus adapted to carry out the method are described, as are other aspects.

Model-based methods and apparatus for classifying an interferent in specimens

Growing community of inventors

Princeton, NJ, United States of America

JinHyeong Park