Abstract: One or more example embodiments provides a system and a method for differentiating a tissue of interest from another part of a medical scanner image, in particular pectoral muscle tissue from breast tissue in an X-ray mammography image. The method comprises providing a medical scanner image; inputting input data into a trained artificial neural network, the input data being based on the provided medical scanner image; generating, by the trained artificial neural network, output data based on the input data, the output data indicating a one-dimensional borderline between at least a part of the tissue of interest and the at least one other part of the medical scanner image; and outputting an output signal comprising or based on the generated output data.

Abstract: Cyclen based compounds of general formula (I) are disclosed. X is nitrogen and Y, Z are —CH—, or X, Z are —CH— and Y is nitrogen, or X, Y are —CH— and Z is nitrogen. R1 is independently selected from H; COOH; benzyloxycarbonyl; fluorenylmethyloxycarbonyl; tert-butoxycarbonyl; methylcarbonyl; trifluoromethylcarbonyl; benzyl; triphenylmethyl; tosyl; mesyl; benzyloxymethyl; phenylsulfonyl; ethoxycarbonyl; 2,2,2-trichloroethyloxycarbonyl; methoxycarbonyl; methoxymethyloxycarbonyl; R2 is selected from H; methylcarbonyl; tert-butyldimethylsilyl; (C1-C4)alkyl; R3 is independently selected from H; (C1-C6)alkyl.

Abstract: A method and system are for identification of at least one medical reference image. An embodiment of the method includes providing a medical representation image based on a current examination image depicting a body part of a first patient; defining a region of interest in the medical representation image; generating a feature signature, at least for the region of interest; comparing the medical representation image with a plurality of medical images of at least one second patient stored in a medical image database, based on the feature signature generated; and identifying at least one medical image in the medical image database as the at least one medical reference image, the at least one medical reference image providing a similarity degree to the medical representation image above a threshold. In an embodiment, the generating is performed using a trained machine-learning algorithm.

Abstract: The invention relates to a method for imaging a three-dimensional object to be examined. According to said method, a three-dimensional parameterized area is determined which is in conformity with an anatomic structure of the three-dimensional object to be examined. The three-dimensional parameterized area is imaged onto a two-dimensional parameterized area. The three-dimensional object to be examined is represented by imaging pixels that are associated with the three-dimensional parameterized area onto the two-dimensional parameterized area. The invention further relates to a method for determining a camera position in a three-dimensional image recording of an object to be examined. The invention also relates to a method for representing a section of an object to be examined. The invention finally relates to a device for imaging a three-dimensional object to be examined.

Abstract: A method and apparatus for deep learning based automatic bone removal in medical images, such as computed tomography angiography (CTA) volumes, is disclosed. Bone structures are segmented in a 3D medical image of a patient by classifying voxels of the 3D medical image as bone or non-bone voxels using a deep neural network trained for bone segmentation. A 3D visualization of non-bone structures in the 3D medical image is generated by removing voxels classified as bone voxels from a 3D visualization of the 3D medical image.

Abstract: A method and system are for identification of at least one medical reference image. An embodiment of the method includes providing a medical representation image based on a current examination image depicting a body part of a first patient; defining a region of interest in the medical representation image; generating a feature signature, at least for the region of interest; comparing the medical representation image with a plurality of medical images of at least one second patient stored in a medical image database, based on the feature signature generated; and identifying at least one medical image in the medical image database as the at least one medical reference image, the at least one medical reference image providing a similarity degree to the medical representation image above a threshold. In an embodiment, the generating is performed using a trained machine-learning algorithm.

Abstract: A computer-implemented method for generating contours of anatomy based on user click points includes a computer displaying an image comprising an anatomical structure and receiving a first user selection of a first click point at a first position on an outward facing edge of the anatomical structure. The computer applies a contour inference algorithm to generate an inferred contour around the outward facing edge based on the first position. Following generation of the inferred contour, the computer receives a second user selection of a second click point at a second position on the image. Then, the computer creates a visual indicator on a segment of the inferred contour between the first position and the second position as indicative of the user's confirmation of accuracy of the segment.

Abstract: A computer-implemented method for generating contours of anatomy based on user click points includes a computer displaying an image comprising an anatomical structure and receiving a first user selection of a first click point at a first position on an outward facing edge of the anatomical structure. The computer applies a contour inference algorithm to generate an inferred contour around the outward facing edge based on the first position. Following generation of the inferred contour, the computer receives a second user selection of a second click point at a second position on the image. Then, the computer creates a visual indicator on a segment of the inferred contour between the first position and the second position as indicative of the user's confirmation of accuracy of the segment.

Abstract: In a method for providing suitable selection connection structures in a graphics programming language for a selection module, first connection structures of at least one first module are determined in a computer, the first connection structures and the first module being contained in graphics training program sections. A selection module and an existing connection structure of the selection module are also received in the computer, the selection module being contained in a graphics computer program, and the existing connection structure designating the connection structure of the selection module in the graphics computer program. Selection connection structures are also chosen from the volume of first connection structures, wherein each of the selection connection structures is a connection structure of the selection module, and wherein each of the selection connection structures comprises the existing connection structure of the selection module.

Abstract: A method and apparatus for deep learning based automatic bone removal in medical images, such as computed tomography angiography (CTA) volumes, is disclosed. Bone structures are segmented in a 3D medical image of a patient by classifying voxels of the 3D medical image as bone or non-bone voxels using a deep neural network trained for bone segmentation. A 3D visualization of non-bone structures in the 3D medical image is generated by removing voxels classified as bone voxels from a 3D visualization of the 3D medical image.

Abstract: A method is for segmenting an object in a medical image with a plurality of iteration steps. In an embodiment of the method, each iteration step includes generating a plurality of patches, a portion of the input image and a patch location being assigned to each patch, the patch location being indicative of the location of the portion of the input image relative to the input image. For each patch of the plurality of patches, the method includes determining a vote location based on the portion of the input image assigned to that patch and determining a target location based on the vote location and the patch location assigned to that patch. Finally, in an embodiment the method includes generating a vote map, each patch of the plurality of patches contributing to a pixel value at the target location of the patch in the vote map.

Abstract: The invention relates to a method for imaging a three-dimensional object to be examined. According to said method, a three-dimensional parameterized area is determined which is in conformity with an anatomic structure of the three-dimensional object to be examined. The three-dimensional parameterized area is imaged onto a two-dimensional parameterized area. The three-dimensional object to be examined is represented by imaging pixels that are associated with the three-dimensional parameterized area onto the two-dimensional parameterized area. The invention further relates to a method for determining a camera position in a three-dimensional image recording of an object to be examined. The invention also relates to a method for representing a section of an object to be examined. The invention finally relates to a device for imaging a three-dimensional object to be examined.

Abstract: A method and system for approximating a deep neural network for anatomical object detection is discloses. A deep neural network is trained to detect an anatomical object in medical images. An approximation of the trained deep neural network is calculated that reduces the computational complexity of the trained deep neural network. The anatomical object is detected in an input medical image of a patient using the approximation of the trained deep neural network.

Abstract: Systems and methods for segmenting a structure of interest in medical imaging data include generating a binary mask highlighting structures in medical imaging data, the highlighted structures comprising a connected component including a structure of interest. A probability map is computed by classifying voxels in the highlighted structures using a trained classifier. A plurality of detaching operations is performed on the highlighted structures to split the connected component into a plurality of detached connected components. An optimal detaching parameter is determined representing a number of the detaching operations. A detached connected component resulting from performing the number of detaching operations corresponding to the optimal detaching parameter is classified as the structure of interest based on the probability map and the trained classifier.

Abstract: A visualization method is provided that allows for the unfolding of a human skeleton from a medical image scan and providing increased efficiency for interacting with the image scan and whole body bone reading from such scans. That is, a full head-to-toe unfolded skeleton view (e.g., a 2D unfolded view) is realized for improved visualization and diagnostic capabilities.

Abstract: A method and system for approximating a deep neural network for anatomical object detection is discloses. A deep neural network is trained to detect an anatomical object in medical images. An approximation of the trained deep neural network is calculated that reduces the computational complexity of the trained deep neural network. The anatomical object is detected in an input medical image of a patient using the approximation of the trained deep neural network.

Abstract: Systems and methods for segmenting a structure of interest in medical imaging data include generating a binary mask highlighting structures in medical imaging data, the highlighted structures comprising a connected component including a structure of interest. A probability map is computed by classifying voxels in the highlighted structures using a trained classifier. A plurality of detaching operations is performed on the highlighted structures to split the connected component into a plurality of detached connected components. An optimal detaching parameter is determined representing a number of the detaching operations. A detached connected component resulting from performing the number of detaching operations corresponding to the optimal detaching parameter is classified as the structure of interest based on the probability map and the trained classifier.

Abstract: A method is disclosed for determining a boundary surface network of a tubular object. An ordered series of contours is first supplied on the basis of image data in a source space. A transformation function is created for at least two consecutive contours in the series, and a unit space segment object is created in the unit space on the basis of the consecutive contours. A local signed distance function is determined in the unit space. In addition, a relative positional information of a query point is determined in the source space from a surface of a segment object in the source space, the segment object being based on the consecutive contours, on the basis of the local signed distance function in the unit space and using the transformation function. Finally, the boundary surface network is created on the basis of the relative positional information that has been determined.

Abstract: A visualization method is provided that allows for the unfolding of a human skeleton from a medical image scan and providing increased efficiency for interacting with the image scan and whole body bone reading from such scans. That is, a full head-to-toe unfolded skeleton view (e.g., a 2D unfolded view) is realized for improved visualization and diagnostic capabilities.

Abstract: A method is disclosed for determining a boundary surface network of the tubular object. In an embodiment, a representation of the tubular object is initially provided on the basis of image data and local dimension information is provided for points of the representation. A subdivided division structure presentation of the tubular object with division cells is then created, which based on the local dimension information, including a different spatial extent. Finally a boundary surface network is derived on the basis of the division structure presentation. Also described are a boundary surface network determination system and a division structure determination system for performing such a method.