Abstract: A method to detect a defect on a lithographic sample includes the following steps: detection light and a detector having at least one sensor pixel are provided. Further, a detection pattern is provided causing a light structure of the detection light being structured at least along one dimension (1D, x). The detection pattern is aligned such that the detector is aligned normal to an extension (xy) of the light structure. Further, a complimentary pattern is provided having a 1D structure which is complimentary to that of the detection pattern. The sample is moved relative to the detection pattern while gathering the detection light on the detector. Further, a reference sample without defects or with negligible defects is provided. The reference sample also is moved relative to the detection pattern while gathering the detection light on the detector. A defect (S1 to S4) localization on the sample is decoded by correlation using the complementary pattern.

Abstract: Improvements in image quality can be obtained by leveraging non-linear neural-network based feature recovery across multiple imaging modalities. Input imaging data is acquired, including first imaging data and second imaging data acquired using different imaging modalities. The first imaging data can have lower image quality and a larger FOV than the second imaging data. The first and second imaging data can be aligned and the aligned regions can be used to train a neural network to minimize the difference between the second imaging data and output data processed from the overlapping portion of the first imaging data. Once trained, the neural network can be used to generate improved image quality output from all of the first imaging data.

Abstract: Improved (e.g., high-throughput, low-noise, and/or low-artifact) X-ray Microscopy images are achieved using a deep neural network trained via an accessible workflow. The workflow involves selection of a desired improvement factor (x), which is used to automatically partition supplied data into two or more subsets for neural network training. The neural network is trained by generating reconstructed volumes for each of the subsets. The neural network can be trained to take projection images or reconstructed volumes as input and output improved projection images or improved reconstructed volumes as output, respectively. Once trained, the neural network can be applied to the training data and/or subsequent data—optionally collected at a higher throughput—to ultimately achieve improved de-noising and/or other artifact reduction in the reconstructed volume.

Abstract: The generation of a 3D reconstruction of an object is disclosed, which includes illuminating the object, capturing image data in relation to the object, and calculating the 3D reconstruction of the object from the image data. The image data contains first image data and second image data, wherein the first image data are captured when the object is illuminated with illumination light, at least some of which, in relation to an object imaging beam path, is reflected light which illuminates the object, wherein the second image data are captured from different recording directions when the object is illuminated with illumination light, at least some of which is guided in the object imaging beam path, and wherein the 3D reconstruction of the object is calculated from the first image data and the second image data.

Abstract: Stereoscopy in which at least one image of a scene is recorded from a first viewing angle, and at least one image of the scene is recorded from a second viewing angle. The scene is recorded multiple times from the first viewing angle. A first combination image is obtained from the various images recorded from the first viewing angle, said combination image according to the stipulation of a comparison algorithm having smaller differences in relation to at least one image also recorded from the second viewing angle or smaller differences in relation to a second combination image obtained from the images from the second viewing angle than each individual image recorded from the first viewing angle. An image of the scene with the depth information is obtained from the first combination image and at least one image from the second viewing angle or the second combination image.

Abstract: A method for operating a mobile terminal including an image recording device includes capturing at least one recording of a scene and checking at least one of optionally several recordings checked for the presence of a light source. Subsequently, a position of the light source is determined relative to an optical center and a shape, an intensity and/or a color is determined for the light source. An algorithm trained in relation to imaging properties of a given optical unit is then used to generate a flare image of a lens flare for this light source and the given optical unit using the position of the light source. The flare image is subsequently combined with the recording to form a combination image and the combination image is stored in a memory component.

Abstract: Improved (e.g., high-throughput, low-noise, and/or low-artifact) X-ray Microscopy images are achieved using a deep neural network trained via an accessible workflow. The workflow involves selection of a desired improvement factor (x), which is used to automatically partition supplied data into two or more subsets for neural network training. The neural network is trained by generating reconstructed volumes for each of the subsets. The neural network can be trained to take projection images or reconstructed volumes as input and output improved projection images or improved reconstructed volumes as output, respectively. Once trained, the neural network can be applied to the training data and/or subsequent data—optionally collected at a higher throughput—to ultimately achieve improved de-noising and/or other artifact reduction in the reconstructed volume.

Abstract: System/Method/Device for labelling images in an automated manner to satisfy a performance of a different algorithm and then applying active learning to learn a deep learning model which would enable ‘real-time’ operation of quality assessment and with high accuracy.

Abstract: A collision avoidance system and method for an x-ray CT microscope processes image data of an object at different angles and generates a model of the object. This model is then used to configure the microscope for operation and possibly avoid collisions between the microscope and the object.

Abstract: A 3D calibration body for spatial calibration of an optical imaging system includes a transparent body and calibration marks embedded in a volume of the transparent body. At least some of the calibration marks are selectively activatable and deactivatable, wherein an activated calibration mark is visible in the visible spectral range and a deactivated calibration mark is not visible in the visible spectral range.

Abstract: A System/Method/Device for segmenting the choroid-scleral layer from an optical coherent tomography (OCT) volume scan. The present system uses a deep learning machine model based on a neural network that include multiple convolution layers, but no deconvolution layers. Rather, the present neural network is based on a novel architecture based on the discrete cosine transform.

Abstract: An x-ray microscopy method that obtains a classification of different particles by distinguishing between different material phases through a combination of image processing involving morphological edge enhancement and possibly resolved absorption contrast differences between the phases along with optional wavelet filtering.

Abstract: Methods and devices for additive manufacturing of workpieces are provided. For analysis during production, a test is carried out using a selected test method. The test results are compared with simulated test results derived during a simulation of the manufacturing and testing. The test may use one or more of a laser ultrasound test unit, an electronic laser speckle interferometry test unit, an infrared thermography test unit, or an x-ray test unit.

Abstract: Improved (e.g., high-throughput, low-noise, and/or low-artifact) X-ray Microscopy images are achieved using a deep neural network trained via an accessible workflow. The workflow involves selection of a desired improvement factor (x), which is used to automatically partition supplied data into two or more subsets for neural network training. The neural network is trained by generating reconstructed volumes for each of the subsets. The neural network can be trained to take projection images or reconstructed volumes as input and output improved projection images or improved reconstructed volumes as output, respectively. Once trained, the neural network can be applied to the training data and/or subsequent data—optionally collected at a higher throughput—to ultimately achieve improved de-noising and/or other artifact reduction in the reconstructed volume.

Abstract: A system for determining the position of a movable object in space includes a marker which is to be applied to the object. The marker has a surface which is subdivided into a plurality of individual fields. The fields each have a statistical noise pattern. The system also includes an image capture unit which is remote from the object and is arranged to capture an image of the marker. The system further includes an image evaluation unit which stores a reference image of the noise patterns and is designed to locate at least one of the fields in the currently captured image of the marker by comparison with the reference image in order to determine a current position of the marker in space. There are corresponding methods for determining a position the object.

Abstract: Stereoscopy in which at least one image of a scene is recorded from a first viewing angle, and at least one image of the scene is recorded from a second viewing angle. The scene is recorded multiple times from the first viewing angle. A first combination image is obtained from the various images recorded from the first viewing angle, said combination image according to the stipulation of a comparison algorithm having smaller differences in relation to at least one image also recorded from the second viewing angle or smaller differences in relation to a second combination image obtained from the images from the second viewing angle than each individual image recorded from the first viewing angle. An image of the scene with the depth information is obtained from the first combination image and at least one image from the second viewing angle or the second combination image.

Abstract: Various approaches in which an image-recording parameter is varied between a plurality of images of an object and a stereo image pair is displayed on the basis of the images recorded thus are described. Here, in particular, the image-recording parameter can be a focal plane or an illumination direction.

Abstract: A method to detect a defect on a lithographic sample includes the following steps: detection light and a detector having at least one sensor pixel are provided. Further, a detection pattern is provided causing a light structure of the detection light being structured at least along one dimension (1D, x). The detection pattern is aligned such that the detector is aligned normal to an extension (xy) of the light structure. Further, a complimentary pattern is provided having a 1D structure which is complimentary to that of the detection pattern. The sample is moved relative to the detection pattern while gathering the detection light on the detector. Further, a reference sample without defects or with negligible defects is provided. The reference sample also is moved relative to the detection pattern while gathering the detection light on the detector. A defect (s1 to s4) localization on the sample is decoded by correlation using the complementary pattern.

Abstract: A coordinate measuring machine has a measurement head having a point measurement device which measures first coordinates of only a single point on the surface of a workpiece at a given time. An area measurement device records images of a reference surface. A displacement device displaces the measurement head and/or the workpiece such that they assume different relative positions with respect to one another. An evaluation device calculates a shift between images that the area measurement device has recorded of the reference surface at different times at different relative positions, with a stitching algorithm. Based on this, second coordinates of the measurement head, which are defined relative to the reference surface, are determined. By linking the first coordinates with the second coordinates, third coordinates are determined, which define the points on the surface of the workpiece measured by the point measurement device relative to the reference surface.

Abstract: An ophthalmic imaging system provides an automatic focus mechanism based on the difference of consecutive scan lines. The system also provides of user selection of a focus point within a fundus image. A neural network automatically identifies the optic nerve head in an FA or ICGA image, which may be used to determine fixation angle. The system also provides additional scan tables for multiple imaging modalities to accommodate photophobia patients and multi-spectrum imaging options.