Siemens Healthcare Gmbh

Siemens Medical Solutions Usa, Inc.

Ahmet Tuysuzoglu

Terrence Chen

Tommaso Mansi

Shun P Miao

Dorin Comaniciu

Shanhui Sun

Tiziano Passerini

Bin Lou

Puneet Sharma

Shaohua Kevin Zhou

Ali Kamen

Yefeng Zheng

Ankur Kapoor

Rui Liao

Berthold Kiefer

Günter Lauritsch

Stefan Kluckner

Robert Grimm

Alexander Brost

Heinrich Von Busch

Mamadou Diallo

Francisco Pereira

Bibo Shi

David Jean Winkel

Tongbai Meng

Guenter Lauritsch

Xin Yu

Nicolas Von Roden

Yuan Liu

Kareem Abdelrahman

A method and system for deep learning based cardiac electrophysiological model personalization is disclosed. Electrophysiological measurements of a patient, such as an ECG trace, are received. A computational cardiac electrophysiology model is personalized by calculating patient-specific values for a parameter of the computational cardiac electrophysiology model based at least on the electrophysiological measurements of the patient using a trained deep neural network (DNN). The parameter of the computational cardiac electrophysiology model corresponds to a spatially varying electrical cardiac tissue property.

System and method for deep learning based cardiac electrophysiology model personalization

For non-invasive EP mapping, a sparse number of electrodes (e.g., 10 in a typical 12-lead ECG exam setting) are used to generate an EP map without requiring preoperative 3D image data (e.g. MR or CT). An imager (e.g., a depth camera) captures the surface of the patient and may be used to localize electrodes in any positioning on the patient. Two-dimensional (2D) x-rays, which are commonly available, and the surface of the patient are used to segment the heart of the patient. The EP map is then generated from the surface, heart segmentation, and measurements from the electrodes.

Non-invasive electrophysiology mapping based on affordable electrocardiogram hardware and imaging

Systems and methods are provided for classifying an abnormality in a medical image. An input medical image depicting a lesion is received. The lesion is localized in the input medical image using a trained localization network to generate a localization map. The lesion is classified based on the input medical image and the localization map using a trained classification network. The classification of the lesion is output. The trained localization network and the trained classification network are jointly trained.

Localization and classification of abnormalities in medical images

A computer-implemented method for decoding brain imaging data of individual subjects by using additional imaging data from other subjects includes receiving a plurality of functional Magnetic Resonance Imaging (fMRI) datasets corresponding to a plurality of subjects. Each fMRI dataset corresponds to a distinct subject and comprises brain activation patterns resulting from presentation of a plurality of stimuli to the distinct subject. A group dimensionality reduction (GDR) technique is applied to the example fMRI datasets to yield a low-dimensional space of response variables shared by the plurality of subjects. A model is trained to predict a set of target variables based on the low-dimensional space of response variables shared by all subjects, wherein the set of target variables comprise one or more characteristics of the plurality of stimuli.

Decoding from brain imaging data of individual subjects by using additional imaging data from other subjects

Methods and apparatus for cross-domain medical image analysis and cross-domain medical image synthesis using deep image-to-image networks and adversarial networks are disclosed. In a method for cross-domain medical image analysis a medical image of a patient from a first domain is received. The medical image is input to a first encoder of a cross-domain deep image-to-image network (DI2IN) that includes the first encoder for the first domain, a second encoder for a second domain, and a decoder. The first encoder converts the medical image to a feature map and the decoder generates an output image that provides a result of a medical image analysis task from the feature map. The first encoder and the second encoder are trained together at least in part based on a similarity of feature maps generated by the first encoder from training images from the first domain and feature maps generated by the second encoder from training images from the second domain, and the decoder is trained to generate output images from feature maps generated by the first encoder or the second encoder.

Cross-domain image analysis and cross-domain image synthesis using deep image-to-image networks and adversarial networks

Reconstructing 3-D vessel geometry of a vessel includes: receiving a plurality of 2-D rotational X-ray images of the vessel; extracting vessel centerline points for normal cross sections of each of the plurality of 2-D images; establishing a correspondence of the centerline points from a registration of the centerline points with a computed tomography (CT) 3-D centerline, the registration being an affine or deformable transformation; constructing a 3-D centerline vessel tree skeleton of the vessel from the centerline points of the 2-D images; constructing an initial 3-D vessel surface having a uniform radius normal to the 3-D centerline vessel tree skeleton; defining sample points based sampling on median radii to the 3-D centerline vessel tree skeleton of the initial 3-D vessel surface; and constructing a target 3-D vessel surface by deforming the initial vessel surface using the sample points to provide a reconstructed 3-D vessel geometry of the vessel.

3-D vessel tree surface reconstruction

A method for reconstructing 3-D vessel geometry of a vessel includes: receiving a plurality of 2-D rotational X-ray images of the vessel; extracting vessel centerline points for normal cross sections of each of the plurality of 2-D images; establishing a correspondence of the centerline points; constructing a 3-D centerline vessel tree skeleton of the vessel; constructing an initial 3-D vessel surface having a uniform radius normal to the 3-D centerline vessel tree skeleton; and constructing a target 3-D vessel surface by deforming the initial vessel surface to provide a reconstructed 3-D vessel geometry of the vessel.

3-D vessel tree surface reconstruction method

Growing community of inventors

Franklin Park, NJ, United States of America

Ahmet Tuysuzoglu