Brainlab Ag

Jens Schmaler

Andreas Blumhofer

Stefan Vilsmeier

Andreas Giese

Patrick Hiepe

Bálint Varkuti

Andreas Blumenhofer

Disclosed are computer-implemented methods which encompass determining whether two medical images were taken of the same patient. In a first aspect, this is done by analysing a registration of the two images with one another. The registration may be a direct registration between the two images or an indirect registration, for example via an atlas to which each image is registered. In other aspects, a machine learning algorithm is trained on the basis of image registrations to determine whether the two images were taken of the same patient. The disclosed methods serve the purpose of being able to group medical images together which were taken of the same patient without having to provide or otherwise process data about the identity of the patient.

Determining image similarity by analysing registrations

Disclosed is a computer-implemented of adapting a biomechanical model of an anatomical body part of a patient to a current status of the patient. The method encompasses determination of a currently executed step of a workflow such as a medical intervention, the result of the determination serving as a basis for adapting and/or updating a biomechanical model of an anatomical body part to the corresponding current status of the patient. The determination of the current workflow step may also be used as basis for controlling an imaging device for tracking entities around the patient or for imaging the anatomical body part or acquiring further data or for urging the user to perform a specific action such as acquisition of information using a tracked instrument such as a pointer. The biomechanical model has been generated from atlas data. The data sets which are generated according to the current workflow step may additionally or alternatively serve as a basis for determining the current workflow step and/or adapting the further workflow.

Adapting a biomechanical model for a patient

Disclosed is a computer-implemented method which encompasses registering a tracked imaging device such as a microscope having a known viewing direction and an atlas to a patient space so that a transformation can be established between the atlas space and the reference system for defining positions in images of an anatomical structure of the patient. Labels are associated with certain constituents of the images and are input into a learning algorithm such as a machine learning algorithm, for example a convolutional neural network, together with the medical images and an anatomical vector and for example also the atlas to train the learning algorithm for automatic segmentation of patient images generated with the tracked imaging device. The trained learning algorithm then allows for efficient segmentation and/or labelling of patient images without having to register the patient images to the atlas each time, thereby saving on computational effort.

Medical image analysis using machine learning and an anatomical vector

Disclosed is a computer-implemented method of segmenting a medical patient image using an atlas and relating the segmentation result to a model of possible geometric changes to the segmentation result (e.g. for correcting the position of the segmentation of anatomical structures) which consider for example anatomical limitations. The thus-related segmentation result may be used as a basis for changing and/or correcting the position, shape and/or orientation of at least parts of the segmentation result, e.g. by user interaction. The invention also relates to an atlas data set comprising information such as values of the variables of the model of possible geometric changes in relation to the positions of anatomical structures in the atlas.

Constrained object correction for a segmented image

Using a current workflow step for control of medical data processing

Disclosed is a computer-implemented methods of determining distributions of corrections for correcting the segmentation of medical image data, determining corrections for correcting the segmentation of medical image data, training a learning algorithm for determining a segmentation of a digital medical image, and determining a relation between an image representation of the anatomical body part in an individual medical image and a label to be associated with the image representation of the anatomical body part in the individual medical image using the trained machine learning algorithm. The methods encompass reading a plurality of corrections to image segmentations, wherein the corrections themselves may have been manually generated, transforming these corrections into a reference system which is not patient-specific such as an atlas reference system, conducting a statistical analysis of the correction, and applying the re-transformed result of the statistical analysis to patient images. The result of the statistical analysis may also be used to appropriately train a machine learning algorithm for automatic segmentation of patient images. The application of such a trained machine learning algorithm is also part of this disclosure.

Correcting segmentation of medical images using a statistical analysis of historic corrections

Disclosed is a computer-implemented method for generating an anonymized medical image of an anatomical body part of a patient, a corresponding computer program, a program storage medium storing such a program and a computer for executing the program, as well as a medical system comprising an electronic data storage device and the aforementioned computer. The disclosed method encompasses establishing a mapping from a patient image onto an atlas, changing that mapping, and applying the inverse of the changed mapping to the atlas in order to transform image content from the atlas to the patient image in order to achieve a deformed and thereby anonymised appearance of the patient image.

Anonymisation of medical patient images using an atlas

The present invention relates to a medical data processing method of transforming a representation of an anatomical structure of a patient in a first imaging modality into a representation of the anatomical structure in a second, other imaging modality, the method being constituted to be executed by a computer and comprising the following steps: acquiring first modality image data describing the first modality medical image containing the representation of the anatomical structure in the first imaging modality; acquiring atlas data describing a first modality atlas image describing a general structure of the anatomical structure in the first imaging modality, the atlas data containing information about the representation of the general structure in the second imaging modality; determining, based on the first modality image data and the atlas data, a first matching transformation between the first modality medical image and the first modality atlas image; determining, based on the first matching transformation and the first modality atlas image and the information about the representation of the general structure in the second imaging modality second modality, a second modality image representation of the first modality medical image.

Growing community of inventors

Mering, Germany

Jens Schmaler