Abstract: A method analyzes optical quality of a glazing region using images of a reference pattern. A first reference image of a first fixed reference pattern is provided or acquired without the glazing, and a first distorted image of the first reference pattern is acquired with the glazing positioned between an optical image acquisition device and the first reference pattern. A second reference image of a reference pattern is provided or acquired without the glazing, and a second distorted image of the reference pattern is acquired with the glazing positioned between the optical image acquisition device and the reference pattern, the reference pattern being located at a distance different from the first reference pattern. Image points are generated on the reference and distorted images, first and second apparent displacement fields are calculated, and a simulated apparent displacement field is determined for a third distance different from the first and second distances.

Abstract: Provided is a method for processing sample data, an electronic device and a storage medium, relating to the field of artificial intelligence technology, and especially to the fields of image processing and computer vision. The method includes: obtaining a plurality of first working images of a spinning box; for each first working image in the plurality of first working images, determining the first working image as a defect image of the spinning box when a difference between the first working image and a normal image of the spinning box meets a first condition; and obtaining a first sample data set based on all defect images in the plurality of first working images.

Abstract: A method and system for training a machine learning model configured to perform characterization of components in a material stream with a plurality of unknown components. A training reward associated with each unknown component within the plurality of unknown components in the material stream is determined, based on which at least one unknown component is physically isolated from the material stream by means of a separator unit, wherein the separator unit is configured to move the selected unknown component to a separate accessible compartment. The isolated at least one unknown component is analyzed for determining the ground truth label thereof, wherein the determined ground truth is used for training an incremental version of the machine learning model.

Abstract: A medical imaging system includes a data storage resource to store probability data representing probability information for a location of a feature of interest in an anatomical region. The system further includes processing circuitry to receive medical image data representing a medical image of the anatomical region, retrieve the probability data from the data storage resource, process the medical image data to perform an image augmentation operation on the received medical image for a feature of interest based on the probability information for the location of the feature of interest in the anatomical region.

Abstract: Provided are a method and system for automatically estimating a mammary gland volume based on a mammary gland magnetic resonance imaging (MRI) image, to automatically segment a mammary gland region in a three-dimensional MRI image of a subject's chest by means of morphological prior knowledge of a breast MRI image and a computer image processing technology, calculate a voxel volume of a segmented part and an actual volume, and guide a doctor in selecting an appropriate fake breast mold. Compared with manual framing of the mammary gland region in conventional practice, the present disclosure can automatically complete marking and volume calculation of the mammary gland region, and greatly improve working efficiency.

Abstract: This application relates to the field of image annotation and recognition in the field of artificial intelligence technologies, and in particular, to a data annotation method. The method includes: using at least two different classification models; pretraining one of the classification models as an initial classification model, and annotating a label for data in a to-be-annotated source dataset as initial data by using the pretrained classification model; and controlling the classification models to perform alternating training and data annotation a quantity of times. Operations of current training and current data annotation include: obtaining data that is re-annotated with a label by a previously trained classification model, selecting a first part of the data to train a current classification model, and re-annotating, by the trained current classification model, a label for a second part of data that is not selected.

Abstract: To achieve faster search processing in the search processing of a similar image, based on a feature value of each of a plurality of keypoints of a human body included in an image, the present invention provides an image processing system 10 including: a target image acquisition unit 11 that acquires a target image; a skeleton structure detection unit 12 that performs processing of detecting a keypoint of a human body included in the target image; a first verification unit 13 that extracts a first reference image whose relationship with the target image satisfies a first extraction condition, from among a plurality of reference images, based on the detected keypoint; and a second verification unit 14 that extracts a second reference image whose relationship with the target image satisfies a second extraction condition, from among the first reference images, based on the detected keypoint.

Abstract: A component mounting system includes multiple production facilities installed in a component mounting line for mounting a component on a board, an imaging section installed in the component mounting line and configured to image the multiple production facilities and one or multiple workers disposed in the component mounting line, and an image processing section configured to perform image processing on an image captured by the imaging section in a predetermined production period in which a product is produced in the component mounting line. The image processing section is configured to visualize and display, for at least one of the one or multiple workers, a work status of the worker for at least one of the multiple production facilities in each of the multiple points in time or multiple periods set in advance in the production period.

Abstract: A target estimation image which is an image of a target at a time point t+w·u obtained by spinning the object in an object image at a time point t obtained from an input video of a plurality of frames in time-series by w unit time on the basis of a hypothesis of spin state and an object image at the time point t+w·u obtained from the input video are used, to estimate the spin state of the object by selecting a hypothesis of spin state and w in which likelihood of the target estimation image becomes high from among a plurality of hypotheses of spin states and a plurality of w whose absolute values are two or more, wherein an absolute value of w is an integer of 2 or more and u is a unit time.

Abstract: Method for characterizing biological particles of a sample, the characterization aiming to determine a property of each particle, the method comprising: a) measuring a characteristic of at least one particle of the sample; b) processing the characteristic of the or each particle using an artificial intelligence algorithm; c) based on the processing, characterizing the particles of the sample, so as to assign a class to each particle, each class being representative of the property of the particle; the method being characterized in that the artificial intelligence algorithm has been trained beforehand by learning from labels, based on training samples, each training sample being assigned a proportion of cells in each class, such that the training is performed on the basis of the proportions respectively assigned to each training sample.

Abstract: A method includes: obtaining a first image of an object including a surface having a non-flat shape; identifying a region corresponding to the surface as a region of interest by applying the first image to a first artificial intelligence model; obtaining data about a three-dimensional (3D) shape type of the object by applying the first image to a second AI model; obtaining a set of values of a 3D parameter related to the object, the surface, or the first camera, based on the region and the data; estimating the non-flat shape of the surface, based on the set of values of the 3D parameter; and obtaining a flat surface image in which the non-flat shape of the surface is flattened, by performing a perspective transformation on the surface.

Abstract: A monitoring device includes an acquisition unit configured to acquire a captured image of a display panel of a control device configured to control an analyzer, an image storage unit configured to store the captured image, and a state determination unit configured to determine a state of the analyzer based on the captured image.

Abstract: A perception system comprises a segmenting neural network and an uncertainty detector. The segmenting neural network is configured and trained for segmentation of an input image pixel matrix to generate a segment composed of elements that correspond the pixels of the input image pixel matrix. Each element of the segment map is assigned to one of a plurality of object classes the segmenting neural network is trained for by way of class prediction. Elements being assigned to the same object class form a segment of the segment map. The uncertainty detector is configured to generate an uncertainty score map that is composed of elements that correspond the pixels of the input image pixel matrix. Each element of the uncertainty map has an uncertainty score that is determined by the uncertainty detector and reflects an amount of uncertainty involved in a class prediction for a corresponding element in the segment map.

Abstract: A method for identifying a jet regulator (1) with improved precision and robustness is provided using a captured photo of a hole pattern (4) of an outlet structure (3) of the jet regulator (1). The captured photo is first compared with different stored hole pattern types (15) in a computer-implemented manner in a first step, and based on an evaluation of the respective degree of a match between the captured photo and each hole pattern type (15), the hole pattern type (15) on which the captured hole pattern (4) is based is ascertained. Subsequently, in another step, the ascertained hole pattern type (15) and the captured hole pattern (4) are examined for deviations in a computer-implemented manner, and identification information which uniquely identifies the jet regulator (1) is ascertained in a computer-implement manner using detected or undetected deviations and is optionally output.

Abstract: Provided is a method of performing registration between a plurality of images of different dimensions. The method may be performed by one or more processors of a computing device, and may include primarily positioning a three-dimensional (3D) image by identifying a first landmark in the 3D image, generating a registration image by registration between the primarily positioned 3D image and a two-dimensional (2D) image, and secondarily positioning the registration image.

Abstract: In order to train a machine learning model such that the machine learning model makes inferences with higher accuracy, a training apparatus includes at least one processor, the at least one processor carries out: an acquiring process of acquiring training images; and a training process of for training the machine learning model with use of the training images, and treats a training image belonging to a subclass different from a subclass to which an image of interest belongs and belonging to a class the same as a class to which the image of interest belongs, as a quasi-positive sample, and a training apparatus belonging to a class different from the class to which the image of interest belong, as a negative sample, to train the machine learning model such that the degree of similarity between features of the image of interest and features of the quasi-positive sample is higher than the degree of similarity between the features of the image of interest and features of the negative sample.

Abstract: Metal artifacts are reduced in x-ray computed tomography (“CT”) images using a suitably trained neural network, such as a convolutional neural network (“CNN”). Virtual metal DATA objects are inserted to either the raw projection data or CT image data (e.g., from pre-procedural CT scans) to generate sets of matching artifact-corrupted and artifact-uncorrupted images, and a CNN, or other neural network, is trained to separate the contribution to each image pixel due to patient anatomy, metal object, or metal object-induced artifact. The contributions from metal object-induced artifacts can then be removed to generate a final, artifact-reduced image.

Abstract: A method for localizing structural connectivity biomarkers in neurological diseases, includes dividing a diffusion magnetic resonance imaging brain volume into a set of connected brain regions; extracting three-dimensional voxels along fiber connections which structurally connect the connected brain regions, wherein the brain regions comprise bundles of neurons; applying a deep neural network to diffusion magnetic resonance imaging features extracted from the three-dimensional voxels for each set of fiber connections which structurally connect brain regions; outputting a disease classification based on applying the deep neural network; and applying multi-instance learning to predict whether each fiber connection is indicative of a healthy brain or a diseased brain.

Abstract: A system, computer readable medium, apparatus and/or method for non-invasive, non-surgical, digital biopsy. The system, computer readable medium, apparatus and/or method accurately predicts benign kidney lesions from cancers in a patient. A processor may receive patient clinical factors, texture analysis of computer-tomographic imaging, and an artificial intelligence learning model. By implementing artificial intelligence, the processor may then predict or determine a probability of kidney cancer in the patient using the patient clinical factors, the texture analysis of computer-tomographic imaging, and the artificial intelligence learning model. Notably, the prediction is performed without needing invasive biopsy surgery and subsequent pathology analysis to arrive at a diagnosis but relies on radiomics metrics.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a storage device, for mapping an electrical grid are disclosed. A method includes sampling multiple locations within a geographic region, executing a detection process for each location, the detection process including applying the set of images for the location as input to a machine learning (ML) model that is trained to identify electrical grid assets depicted within images taken from a combination of different perspectives and obtaining an output from the machine learning model that indicates whether a same electrical grid asset is identified in each of the images of the location. In response to an ML output that indicates a positive identification of the same electrical grid asset being depicted in a particular set of images of a particular location, the method further includes: selecting a number of sublocations within the region, and executing the detection process for each sublocation.