Abstract: A method for processing three-dimensional (3D) point cloud data based on incremental sparse 3D convolution is provided. A computer device obtains 3D point cloud data and forms a training set by processing the 3D point cloud data. The computer device constructs and trains a sparse 3D convolutional network model by inputting the training set. The computer device constructs an incremental sparse 3D convolutional network model by performing incremental replacement of sparse convolutional layers of the trained sparse 3D convolutional network model. The computer device inputs real-time 3D point cloud data into the incremental sparse 3D convolutional network model, and determines an output result as a result of processing the real-time 3D point cloud data. Processing of the 3D point cloud data at least includes 3D semantic segmentation, target detection, 3D classification and video processing.

Abstract: A system for calibrating a driver monitoring system in a vehicle comprises a rearview assembly comprising a housing, a mount secured to a vehicle surface, and a barrel extending between the housing and the mount; an imager disposed in the rearview assembly; a processor in communication with the imager; a memory comprising a file containing information on stationary landmarks proximate the imager; and a controller operable to: receive images of the vehicle's interior and calibrate the imager by identifying a substantially stationary landmark in a first image used during an initial calibration; wherein the controller is configured to re-calibrate the driver monitoring operations based, at least in part, on a detection of the substantially stationary landmark being in a different location in a second image relative to the first image. One of the ends of the barrel is connected with a moveable joint.

Abstract: The disclosure discloses an image processing method and apparatus, and a hardware apparatus. The image processing method includes the following steps: an audio is acquired, and then is preprocessed to obtain audio attribute data at each first time point of the audio; first audio attribute data corresponding to a current time point is acquired; and preset processing is performed on an image to be processed according to the first audio attribute data. According to the image processing method of the embodiments of the disclosure, when performing preset processing on the image to be processed according to the audio attribute data, the image processing can be completed only by setting the relationship between the audio attribute data and the image processing operation, which improves the flexibility and efficiency of image processing.

Abstract: An intelligent understanding apparatus for real-time reconstruction of a large-scale scene light field includes the following. A data obtaining module obtains a 3D instance depth map, and obtain 3D voxels and voxel color information through simultaneous positioning and map generation. The model constructing module constructs and trains a real-time light field reconstruction network model using a ScanNet dataset. The real-time light field reconstruction network model extracts features of the 3D voxels and voxel color information, and obtain a semantic segmentation result and an instance segmentation result. The semantic segmentation module inputs the 3D voxel and voxel color information corresponding to the 3D instance depth map into the trained real-time light field reconstruction network model, and determine an output as a semantic segmentation result and an instance segmentation result corresponding to the 3D instance depth map.

Abstract: A method for cargo counting, a computer equipment, and a storage medium are provided in the disclosure. The method includes the following. Three-dimensional (3D) point cloud data of a set of cargoes within a preset placement region is obtained based on a cargo-counting instruction. Whether the set of cargoes are in a first placement state is determined according to the 3D point cloud data. Based on a determination that the set of cargoes are in the first placement state, a quantity of the set of cargoes is calculated.

Abstract: The present disclosure relates to a method of selecting an accident image by using speed profile analysis, which can sufficiently secure an available capacity of a storage medium, can reduce the amount of transmission data and a fee therefor, and can prevent a loss of unnecessary management expenses, by selecting an actual accident image by using speed profile analysis before and after the occurrence of an impact event and deleting, from the storage medium, an image having a grade determined to have a low accident possibility or changing a state of the image into an overwriteable state or taking measures for preventing the transmission of the image to a cloud server.

Abstract: Based on a measured intensities of first and second light sources while both the first and second light sources are actuated, and a default secondary illuminator output intensity, a computer may determine an adjusted secondary illuminator output intensity and actuate the secondary light source to output light at the adjusted secondary light source output intensity.

Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods for utilizing machine learning models to generate refined depth maps of digital images utilizing digital segmentation masks. In particular, in one or more embodiments, the disclosed systems generate a depth map for a digital image utilizing a depth estimation machine learning model, determine a digital segmentation mask for the digital image, and generate a refined depth map from the depth map and the digital segmentation mask utilizing a depth refinement machine learning model. In some embodiments, the disclosed systems generate first and second intermediate depth maps using the digital segmentation mask and an inverse digital segmentation mask and merger the first and second intermediate depth maps to generate the refined depth map.

Abstract: A candidate face image and attribute information of each imaging attribute item are acquired. For each imaging attribute item, attribute information is extracted from a database that stores attribute information that has a degree of participation in false authentication greater than a participation degree threshold, among attribute information of analyzed imaging attribute items of analyzed master face images that resulted in false authentication in face image authentication. For each imaging attribute item, whether or not the extracted attribute information has a similarity to the acquired attribute information of the each imaging attribute item greater than a similarity threshold is determined. For each imaging attribute item for which the similarity is greater than the similarity threshold, a weight is assigned. A total score is calculated by summing the weights. When the total score satisfies a predetermined threshold value, the candidate face image is determined as inappropriate as a master face image.

Abstract: Systems and methods relate to processing digital-pathology images. More specifically, aspects of the present disclosure are directed to accessing a whole-slide image depicting a slice of specimen, defining a set of tiles within at least part of the whole-slide image, generating one or more artifact prediction metrics by applying artifact detection to each tile of the set of tiles, wherein each of the one or more artifact prediction metrics corresponds to a predicted level of image quality of part or all of the whole-slide image, generating a quality heat map image corresponding to the whole-slide image, wherein the quality heat map image is based on the one or more artifact prediction metrics, and outputting the quality heat map image.

Abstract: A slow-motion video shooting method includes detecting that a preset motion occurs on a main object in a video stream; determining a first video clip in the video stream based on the preset motion; determining a first frame rate based on the preset motion; and processing the first video clip based on the first frame rate to obtain a second video clip, where a play time of the second video clip at a target play frame rate is greater than a collection time of the first video clip, and the first frame rate is greater than the target play frame rate. When a main object for shooting moves, the first video clip, for example, a highlight clip, is obtained from the video stream, so that an electronic device can accurately capture a highlight moment of the main object.

Abstract: A towing assist system for a vehicle includes a marker positioned on a trailer. An imaging device is in connection with the vehicle and configured to capture image data demonstrating a rearward-directed field of view. A controller processes the image data depicting the characteristic marker and detects the characteristic marker in the image data. The controller further identifies a coupler position or coupler height of a coupling interface of the trailer in response to a position of the marker in at least one frame of the image data.

Abstract: The present disclosure relates to a method and an apparatus for measuring motility of ciliated cells in a respiratory tract. The method includes the operations of: acquiring image data including a plurality of frames of respiratory tract organoids; identifying positions of ciliated cells by performing motion-contrast imaging on the image data; when a region of interest (ROI) related to the position of the ciliated cells is selected, measuring a ciliary beat frequency (CBF) related to motility of cilia included in the selected region of interest using cross-correlation between the plurality of frames; and expressing the cilia included in the region of interest in a preset display method on the basis of the range of the measured ciliary beat frequency.

Abstract: A method to verify documentation, such as healthcare documentation, in real-time is described. An image of a first document and demographic information is received from a patient. Optical character recognition (OCR) or QR code, barcode, or machine readable zone (MRZ) reading is performed on the image of the first document and a patient identifier is extracted from the first document. The patient identifier is compared to content in a database to identify a second document related to the first document that shares the patient identifier with the first document. The first document and the second document are verified as being associated with the patient based on the comparison and predefined business rules. OCR errors are identified in the first document and corrected using content from the second document. The method occurs prior to and allows the patient to pay for a healthcare appointment prior to attending the appointment.

Abstract: A method for inspecting a component made of press-hardening steel prior to resistance spot welding of the component includes performing non-destructive testing of the component made of press-hardening steel after hot stamping to determine a plurality of characteristics for the component. The non-destructive testing comprises at least one of image processing, electromagnetic analysis, and elemental analysis of the component. The method includes using a model correlating values of the characteristics to acceptable weld quality or rejected weld quality and predicting acceptable weld quality or rejected weld quality of the component prior to resistance spot welding of the component; resistance spot welding the component if the model predicts acceptable weld quality; and not resistance spot welding the component if the model predicts rejected weld quality.

Abstract: An apparatus for producing constrained medical image data, the apparatus including processing circuitry configured to: receive medical image data that includes or is obtained from scan data representing an anatomical region in which a sub-region is enhanced; predict, using a trained model, mask data from the medical imaging data, wherein the mask data is representative of the anatomical region without enhancement of the sub-region; and predict, using the trained model or a further trained model, subtraction data from the same medical image data, the subtraction data being representative of the same anatomical region, and the processing circuitry being further configured to apply at least one constraint to obtain constrained subtraction data.

Abstract: Systems and techniques for generation and curation of a professional headshot from a set of image data. The systems and techniques images from the set of image data based on characteristics of the representation of the individual within the image. The systems and techniques further include determining a bounding box to define a headshot, the bounding box determined based on guidelines established by heuristics and/or machine learning algorithms trained using data labeled based on heuristics.

Abstract: A classification system configured to obtain class determination data, perform classification on the basis of the class determination data, and check appropriateness of individual candidate classes in a candidate class list obtained via classification, by storing attributes of the candidate classes and checking the candidate classes against the stored attributes. By checking the appropriateness of individual candidate classes via a previously-constructed database in which the attributes of classes are stored, and by making comparisons as to whether the attributes of the respective candidate classes match class determination data and allowing the user to explain the reason for classification, the accuracy and explainability of the classification system can be improved.

Abstract: Embodiments of the present disclosure relate to a method, an electronic device, and a computer program product for image segmentation. The method may be executed by a trained image segmentation model. The method includes: obtaining a to-be-processed image, wherein the to-be-processed image includes objects of a plurality of categories. The method further includes: selecting a to-be-discarded category in the to-be-processed image according to a recall rate of each of the plurality of categories obtained in advance. The method further includes: processing the to-be-processed image based on a plurality of remaining categories in the plurality of categories other than the to-be-discarded category to obtain a segmented image. Through the method, computing resources required for a segmentation processing task can be greatly reduced, a processing amount of image data can be reduced, and an image processing speed can be increased.

Abstract: An electronic device is provided. The electronic device includes a memory, an image sensor including light receiving elements each including at least two sub light receiving elements, and an image signal processor. The image signal processor is configured to obtain images corresponding to light from outside by using the image sensor, the images including at least a raw image, a first sub image, and a second sub image, the first sub image being an image corresponding to light detected by at least one first sub light, the second sub image being an image corresponding to light detected by at least one second sub light, identify a luminance ratio between the first sub image and the second sub image, identify a defect in the raw image, based on the luminance ratio, and perform a function corresponding to a type of the defect.