Abstract: Disclosed are systems and methods for training an active learning system. In one example, a method for training an active learning system includes the steps of generating a point cloud based on a first image captured at a first pose, projecting the point cloud into a second image captured at a second pose to determine an overlap ratio between the point cloud and the second image, and using the second image to train the active learning system based on probabilities associated with the points of the point cloud projected into the second image and the overlap ratio.

Abstract: A method, apparatus, system, and computer program product for detecting inconsistencies in an unconsolidated composite material. Fiber inconsistencies are detected in sections of the fibers input into a composite material manufacturing system in real time during operation of the composite material manufacturing system. Material inconsistencies are detected in the unconsolidated composite material in real time as the unconsolidated composite material is being manufactured by the composite material manufacturing system using the fibers. A determination is made as to whether the material inconsistencies are present in a number of the sections in which the fiber inconsistencies are detected in real time as the unconsolidated composite material is being manufactured by the composite material manufacturing system.

Abstract: A mobile communications device for performing localization and mapping comprising at least one visual sensor for capturing visual sensor data of a local environment, and operative to derive key-point descriptors from the visual sensor data which represent visual features of the local environment, derive concealed representations of the derived key-point descriptors, by applying a current permutation to the derived key-point descriptors, determine a pose of the mobile communications device relative to a point-cloud map representing the local environment, map the derived key-point descriptors onto the point-cloud map, thereby associating each derived key-point descriptor with a position in the point-cloud map, and transmit the concealed representations of the derived key-point descriptors, together with their associated positions in the point-cloud map, to a map server.

Abstract: Embodiments of the present disclosure provide a method, an electronic device, and a computer program product for recognizing object regions in an image. The method includes detecting edges in an image based on pixels in the image. The method further includes segmenting the image into a plurality of semantic regions based on semantics of objects in the image. The method further includes recognizing, based on the edges and the plurality of semantic regions, object regions covering the objects. In this way, the positions of objects including graphic elements in the image and the tags of the objects can be automatically recognized, thus improving the efficiency of subsequent processing of the image.

Abstract: A medical image processing apparatus includes processing circuitry to receive radiation image data for each of a plurality of different channels, wherein the radiation image data for all of the plurality of channels represents a same anatomical region of a same subject; and, for each of a plurality of positions represented in the radiation image data: estimate, based on data values for the position in each of the plurality of channels, material probabilities which indicate a respective probability of each of a plurality of materials existing at the position; and specify a value for at least one rendering parameter at the position based on the material probabilities.

Abstract: A slit lamp microscope of an aspect example includes a scanner and a data processor. The scanner is configured to scan an anterior segment of a subject's eye with slit light to collect a plurality of cross sectional images. The data processor is configured to generate opacity distribution information that represents a distribution of an opaque area in a crystalline lens, based on the plurality of cross sectional images collected by the scanner.

Abstract: An information handling system peripheral camera analyzes captured visual images to detect a face and analyzes the face to determine an upright orientation. Based upon the face upright orientation, the camera defines a visual image orientation to communicate the visual image from the peripheral camera to an information handling system so that the face is communicated in an upright orientation for use in a video conference, such as with an indication of a landscape or portrait image orientation. In one embodiment, an information handling system can perform the face upright orientation analysis and command a camera orientation to the camera so that visual images of the camera are communicated in alignment with the face upright orientation.

Abstract: A system and methods for recreating images from undersampled magnetic resonance imaging (MRI) image data to remove artifacts from the undersampling. Using trained transformer networks, undersampled k-space data can be acquired and the networks can predict additional projected radial MRI k-space data from the undersampled radial MRI k-space data. Images with fewer artifacts can then be generated.