Abstract: Systems, devices, methods, and computer-readable media for image processing by machine learning are provided. A method can include providing, as input to a machine learning (ML) model for image processing, data of a first image, receiving, from the ML model, an estimated radar cross section (RCS) of the first image, receiving, from the ML model, estimated transformation parameters that project the first image to a second image space, and applying the estimated transformation parameters to the estimated RCS to generate an estimated second image in the second image space.

Abstract: A computer, including a processor and a memory, the memory including instructions to be executed by the processor to determine a first measure of pixel values in the first image, acquire a second image, estimate ambient light illuminating the first object based on the second image and modify the first measure of pixel values based on a second measure of pixel values corresponding to estimated ambient light based on the second image. The instructions include further instructions to perform a comparison of the modified first measure of pixel values to a third measure of pixel values determined from a third image of a second object, wherein the third image is previously acquired by illuminating the second object with a second light beam and, when the comparison determines that the first measure is equal to the third measure of pixel values within a tolerance, determine whether the first object and the second object are a same object.

Abstract: A control device, which controls a machine tool capable of alarm stop based on abnormality in machining load, includes: a photographing unit which photographs chips produced as a result of machining of a workpiece; a reference model determination unit which determines in advance a reference model of chips for determining as abnormality in machining, based on images of chips photographed within a predetermined period before alarm stop, in response to an occurrence of alarm stop; and a judgment unit which judges as abnormality occurrence in machining, in a case of a degree of similarity in chips photographed at a predetermined timing in machining later, relative to a reference model of chips that was determined.

Abstract: An electronic device for optically detecting an appearance of a product for defects includes a first camera device, a second camera device, at least one white light source, and at least one red light source. The first camera device is perpendicular to a plane of the product to be detected. When the red light source is activated, the first camera device captures images of edges of the plane of the product, when the white light source is activated, the second camera device captures images of the plane of the product. The electronic device detects the appearance of the plane of the product for defects according to the images captured by the first and second camera devices, the defects can include stair slope errors, scratches, sanding marks, and gas marks.

Abstract: In one embodiment, a method includes accessing a pair of stereo images for a scene, where each image of the pair of stereo images has incomplete pixel information and k channels, stacking the pair of stereo images to form a stacked input image with 2k channels, processing the stacked input image using a machine-learning model to generate a stacked output image with 2k channels, and separating the stacked output image with 2k channels into a pair of reconstructed stereo images for the scene, where each image of the pair of reconstructed stereo images has complete pixel information and k channels.

Abstract: A method is disclosed including: receiving raw image data corresponding to a series of raw images; processing the raw image data with an encoder to generate encoded data, where the encoder is characterized by an input/output transformation that substantially mimics the input/output transformation of one or more retinal cells of a vertebrate retina; and applying a first machine vision algorithm to data generated based at least in part on the encoded data.

Abstract: Methods and systems for incorporating physical documents into a document review workflow involving electronic documents. One or more embodiments detect a presence of a physical document within a field of view of an AR device and map the physical document to an existing electronic document based on visual features of the physical document. Additionally, one or more embodiments determine at least one difference between the physical document and the electronic document and create, for the physical document and the electronic document a shared state mapping including the difference(s). One or more embodiments then apply the difference to the physical document or the electronic document by displaying the difference(s) in an AR layer within the field of view of the AR device or storing the difference(s) in the electronic document.

Abstract: An object recognition device 80 includes a scene determination unit 81, a learning-model selection unit 82, and an object recognition unit 83. The scene determination unit 81 determines, based on information obtained during driving of a vehicle, a scene of the vehicle. The learning-model selection unit 82 selects, in accordance with the determined scene, a learning model to be used for object recognition from two or more learning models. The object recognition unit 83 recognizes, using the selected learning model, an object in an image to be photographed during driving of the vehicle.

Abstract: Provided are a method and an apparatus for detecting pixel defect of optical module, and a device, where the method includes: graying an image obtained by imaging a test binary image by an optical module in a darkroom environment to obtain a first grayscale image; determining a first grayscale area and a second grayscale area corresponding to two gray levels in the test binary image from the first grayscale image; determining a pixel point not matching a grayscale feature of respective grayscale region from the first grayscale area and the second grayscale area respectively as a pixel defect point; and determining the pixel defect of the optical module according to the pixel defect point. The technical solution provided in the present disclosure can detect accurately a pixel defect of an optical module which is advantageous in optimizing the processing technology of the optical module.

Abstract: Quality associated with an interpretation of data captured as unstructured data can be determined. Attributes can be identified within the unstructured data automatically. Subsequently, sentiment associated with each of the attributes can be determined based on the unstructured data. Correctness of the unstructured data, and thus the interpretation, can be assessed based on a comparison of the attribute and associated sentiment with structured data. A quality score can be generated that captures the quality of the data interpretation in terms of correctness and as well as results of another analysis including completeness, among others. Comparison of the quality score to a threshold can dictate whether or not the interpretation is subject to further review.

Abstract: Interfaces and systems are provided for harvesting ground truth from forms to be used in training models based on key-value pairings in the forms and to later use the trained models to identify related key-value pairings in new forms. Initially, forms are identified and clustered to identify a subset of forms to label with the key-value pairings. Users provide input to identify keys to use in labeling and then select/highlight text from forms that are presented concurrently with the keys in order to associate the highlighted text with the key(s) as the corresponding key-value pairing(s). After labeling the forms with the key-value pairings, the key-value pairing data is used as ground truth for training a model to independently identify the key-value pairing(s) in new forms. Once trained, the model is used to identify the key-value pairing(s) in new forms.

Abstract: Methods, systems, and computer-readable storage media for receiving input data including a set of entities of a first type and a set of entities of a second type, providing a set of features based on entities of the first type, the set of features including features expected to be included in entities of the second type, filtering entities of the second type based on the set of features to provide a sub-set of entities of the second type, and generating an output by processing the set of entities of the first type and the sub-set of entities of the second type through a ML model, the output comprising a set of matching pairs, each matching pair in the set of matching pairs comprising an entity of the set of entities of the first type and at least one entity of the sub-set of entities of the second type.

Abstract: A computational methodology for noninvasively assessing the severity of arterial stenosis and predicting the therapeutic outcome of interventional treatment for stenosis assessed as severe, mild, or in between based on patient's CT/MRI imaging data, ultrasound test data, and physio-pathological material properties. The method includes two major parts. The steps in the first part comprise receiving medical data, segmenting the anatomical three-dimensional geometry of the stenosed artery, setting up boundary conditions at inlet and outlets using the ultrasound velocity waveforms together with 3-element WinKessel model, and computing pulsatile pressure waveforms proximal and distal to the existing stenosis for TPI.

Abstract: A radiographic system includes a storage unit configured to store a plurality of imaging methods in association with a plurality of imaging protocols, a determination unit configured to determine whether an imaging method of an imaging protocol associated with an examination order matches at least one of the plurality of imaging methods stored in the storage unit, and an image processing unit configured to apply, in a case where the determination unit determines that the imaging method matches at least one imaging method, an image processing condition of an imaging protocol corresponding to the at least one imaging method to a radiographic image captured based on the imaging protocol associated with the examination order.

Abstract: The present invention relates to an image processing device and a method of framing changes and movements in a video image divided into N×N blocks of pixel positions. The method comprises calculating a first bitmap of the video image by a DCT transform on each of the N×N blocks of pixel positions, assigning a first binary value to the pixel positions of the N×N blocks when more than an amount of change, and a second binary value to the pixel positions of the N×N blocks when less than an amount of change. Calculating a third bitmap by an OR operation between a number of bitmaps representing past time frames of the video image, calculating a fourth bitmap by performing a dilation process of the third bitmap representing the current time frame of the video image, and creating one or more frames identifying area of changes and movements in the video image based on detecting BLOBs (Binary Large Objects) in the fourth bitmaps.

Abstract: An image processing apparatus including a determination unit configured to determine, for each of a plurality of partial areas set in an input image, a priority for executing a first image analysis process, based on an image of the partial area and a corresponding image corresponding to the partial area, the corresponding image being in a past image analyzed before the input image, a selection unit configured to select at least one of partial areas from the plurality of partial areas, based on the determined priority, and a first analysis unit configured to execute the first image analysis process on each of the at least one of partial areas selected.

Abstract: The disclosure provides 3D reconstruction methods and devices. The method includes: obtaining data captured by the camera and data captured by the inertial measurement unit; obtaining a pose of the camera based on the data; obtaining an adjustment value of the pose of the camera and an adjustment value of bias of the inertial measurement unit; updating the pose of the camera based on the adjustment value of the pose of the camera; determining whether the adjustment value of bias of the inertial measurement unit is less than a preset value; in response to the adjustment value of bias of the inertial measurement unit being greater than or equal to the preset value, determining that a current loop for 3-dimensional reconstruction is an error loop; removing the error loop; and constructing a 3-dimensional model for surroundings of the camera based on the updated pose of the camera and remaining loops.

Abstract: A system is configured to perform operations includes accessing a set of model points of a model of an anatomic structure of a patient, the model points being associated with a model space. A set of measured points of the anatomic structure of the patient are collected, the measured points being associated with a patient space. The set of model points are registered to the set of measured points using a first set of initial parameters to generate a first transformation. One or more sets of perturbed initial parameters are generated based on the first set of initial parameters. One or more perturbed registration processes are performed to register the set of model points to the set of measured points using the one or more sets of perturbed initial parameters respectively to generate corresponding perturbed transformations. A registration quality indicator is generated based on the first transformation and the one or more perturbed transformations.

Abstract: A set of pre-operative images may be captured of an anatomical structure using an endoscopic camera. Each captured image is associated with a position and orientation of the camera at the moment of capture using image guided surgery (IGS) techniques. This image data and position data may be used to create a navigation map of captured images. During a surgical procedure on the anatomical structure, a real-time endoscopic view may be captured and displayed to a surgeon. The IGS navigation system may determine the position and orientation of the real-time image; and select an appropriate pre-operative image from the navigation map to display to the surgeon in addition to the real-time image.

Abstract: A method of global movement image stabilization may include calculating a global motion group parameter with respect to first and second sequential frames for two or more elementary 2D motion components of a motion group. Each global motion group parameter may be applied to a motion group vector field that corresponds to the elementary 2D motion component to which the respective global motion group parameter applies to generate a global motion group vector field corresponding to each elementary 2D motion component of the motion group. The global motion group vector fields may be summed pointwise to generate a global motion deformation vector field that provides global motion from the second frame to the first frame. The global deformation vector field may be cumulated with a previous cumulative global deformation vector field that provides global movement from the first frame to one or more previous frames to generate a current cumulative global motion deformation vector field.