Siemens Healthineers Ag

Siemens Healthcare Gmbh

Alexander Preuhs

Michael Manhart

Günter Lauritsch

Sailesh Conjeti

Thomas Redel

Markus Kowarschik

Christian Kaethner

Martin Berger

Christian Huemmer

Sebastian Bauer

Yixing Huang

Lei Wang

Manuela Meier

Mario Orsini

Techniques are described to infer 2-D segmentations of a region of interest using a neural network algorithm. Techniques are described to train the neural network algorithm. The 2-D segmentations are determined based on multiple 2-D projection images. For example, x-ray images can be used as an input.

Processing 2-D projection images using a neural network

A method for evaluation of medical image data comprises: providing medical image data of a patient to be examined; determining, for at least one segment of the medical image data, a respective classification probability value with respect to at least one classification from a list of specified classifications; determining a patient-specific relevance criterion for at least one classification for at least the at least one segment of the medical image data; and determining a clinical relevance of the at least one classification for the at least one segment of the medical image data using the patient-specific relevance criterion, and at least one of based on the classification probability values or based on the at least one segment of the medical image data.

Method and apparatus for the evaluation of medical image data

A method includes receiving image data of an examination object. A first temporary data record is created by applying a first correction to the image data. A further temporary data record is created by applying a further correction to the image data. The further correction at least partially corresponds to the first correction. A trained function is applied to input data that is based on the first temporary data record and the further temporary data record. A parameter of the trained function is based on an image quality metric. It is determined whether the first temporary data record has a higher image quality compared with the further temporary data record. When a result is positive, the first temporary data record is provided as the corrected medical image data. When the result is negative, the further temporary data record is provided as the image data, and part of the method is repeated.

Provision of corrected medical image data

In a computer-implemented method of training a reconstruction neural network algorithm used to reconstruct a Magnetic Resonance Imaging (MRI) image, a prediction of training MRI image is determined based on training MRI raw data and using the reconstruction neural network algorithm. A prediction of a presence or absence of the object in the training MRI image is determined based on the prediction of the training MRI image and using an object-detection algorithm. A loss value is determined based on a first difference between the ground truth of the training MRI image and the prediction of the training MRI image, and further based on a second difference between the ground truth of the presence or absence of the object and the prediction of the presence or absence of the object. Weights of the reconstruction neural network algorithm are adjusted based on the loss value and using a training process.

Task-specific training of reconstruction neural network algorithm for magnetic resonance imaging reconstruction

For creation of a movement-compensated image reconstruction in an x-ray based imaging method an optimal object movement trajectory is determined by application of a trained algorithm to an optimal latent vector and the movement-compensated image reconstruction is created depending on projection images, with the assumption that the object would have moved in accordance with the optimal object movement trajectory while the imaging method was being carried out. The optimal latent vector is determined by a predetermined cost function being minimized using the latent vector as an optimization parameter, wherein for each optimization step with a current latent vector, a current object movement trajectory is determined by application of a trained algorithm to the current latent vector, and a current value of the cost function is computed depending on the plurality of projection images, with the assumption that the object would have moved in accordance with the current object movement trajectory while the imaging method was being carried out.

Movement-compensated image reconstruction for an x-ray based imaging method

Systems and methods are provided for reconstructing a three-dimensional result image data set from computed tomography from a plurality of two-dimensional images that create an image of an object undergoing examination from a particular imaging angle, The imaging angles of all the images lie within a restricted angular range. A three-dimensional artifact-reduced image data set is provided based on the two-dimensional images using an algorithm for reducing artifacts that are the result of a restriction of the angular range. The result image data set is reconstructed using a reconstruction algorithm that processes both the artifact-reduced image data set and the two-dimensional images as input data.

Method for reconstructing a three-dimensional image data set

A method is provided for determining corrected acquisition geometries of projection images. The method includes providing a projection image dataset that has a plurality of projection images of an object under examination acquired by an acquisition device in different acquisition geometries. The method further includes determining a provisional acquisition geometry for each of the projection images by a first optimization method by minimizing a first cost function by varying the provisional acquisition geometry, wherein the first cost function is contingent on a plurality of consistency measures determined based on the provisional acquisition geometry for a respective pair of projection images. The method further includes determining the respective corrected acquisition geometry for each of the projection images by a second optimization method by minimizing a second cost function by varying the corrected acquisition geometries, wherein the second cost function is contingent on a measure for an image quality of image data reconstructed based on the projection images and the corrected acquisition geometries, and wherein, in a first iteration act of the second optimization method, the provisional acquisition geometries determined by the first optimization method are used as corrected acquisition geometries.

Method for determining corrected acquisition geometries of projection images

A method and system are provided for at least symbolically reconstructing a reconstruction data set of at least one vessel segment in a vessel tree of a patient. Input data for the reconstruction comprises at least two two-dimensional angiographic projection images taken in different acquisition geometries. At least one first angiographic projection image showing the vessel segment is acquired. An evaluation measure is automatically determined for each first angiographic projection image using three-dimensional preliminary information for the vessel segment. The evaluation measure describes the suitability of the at least one angiographic projection image for reconstructing the reconstruction data set. When a quality criterion evaluating the evaluation measure is not fulfilled, at least one additional acquisition geometry is determined using the three-dimensional preliminary information and/or the evaluation measure. In each additional acquisition geometry, at least one second angiographic projection image is acquired. The reconstruction data set is reconstructed from the at least one second angiographic projection image and/or at least one of the at least one first angiographic projection image fulfilling a suitability criterion evaluating the evaluation measure.

Growing community of inventors

Erlangen, Germany

Alexander Preuhs