Abstract: Disclosed are techniques for enhancing two-dimensional (2D) image capture of subjects (e.g., a physical structure, such as a residential building) to maximize the feature correspondences available for three-dimensional (3D) model reconstruction. More specifically, disclosed is a computer-vision network configured to provide viewfinder interfaces and analyses to guide the improved capture of an intended subject for specified purposes. Additionally, the computer-vision network can be configured to generate a metric representing a quality of feature correspondences between images of a complete set of images used for reconstructing a 3D model of a physical structure. The computer-vision network can also be configured to generate feedback at or before image capture time to guide improvements to the quality of feature correspondences between a pair of images.

Abstract: A computer-implemented method for analyzing refuse includes operations of receiving sensor data indicating an operational state of a vehicle body component of a refuse collection vehicle (RCV); analyzing the sensor data to detect a presence of a triggering condition based at least partly on a particular operational state of the vehicle body component, as indicated by the sensor data; in response to detecting the triggering condition, accessing image data indicating a physical state of refuse collected by the RCV; providing the image data as input to at least one contaminant detection model trained, using at least one machine learning (ML) algorithm, to output a classification of the image data, the classification indicating a degree of contamination of the refuse; and storing, in a machine-readable medium, the classification of the image data.

Abstract: Techniques are provided to detect and/or replace anomalous pixels. In one example, a method includes receiving an image frame comprising a plurality of pixels having associated pixel values. The method also includes selecting a kernel of the pixels comprising a center pixel and a plurality of neighbor pixels. The method also includes, if the center pixel value exhibits an anomalous pixel condition, calculating a replacement pixel value using a gradient associated with at least a subset of pixel values of the neighbor pixels. Additional methods and systems are also provided.

Abstract: A computational platform and a method for use in a depth calculation process based on information comprised in an image captured by one or more image capturing sensors, wherein the computational platform enables distinguishing between areas included in the captured image that comprise details that are implementable by a matching algorithm and areas that do not have such details, wherein the computational platform comprises at least one processor, configured to select at least one matching window comprised in the captured image for matching a corresponding part included in each image captured by the one or more image capturing devices; calculate a metric based on a respective selected matching window; and calculate a depth map based on the calculated metric associated with the at least one matching window.

Abstract: A plane detecting device according to the present disclosure includes an information acquisition unit, a likelihood acquisition unit, and a plane detector. The information acquisition unit acquires visible image information of a target having a predetermined plane and 3D coordinate information corresponding to the visible image information. The likelihood acquisition unit acquires likelihoods indicating a planarity of the predetermined plane of the target from the visible image information. The plane detector detects the predetermined plane of the target through a robust estimation method by using the 3D coordinate information and the likelihood.

Abstract: Disclosed is method, a user interface, a computer program product and a system for predicting future development of a dental condition of a patient's set of teeth, wherein the method comprises: —obtaining two or more digital 3D representations for the teeth recorded at different points in time; —deriving based on the obtained digital 3D representations a formula expressing the development of the dental condition in terms of at least one parameter as a function of time; and—predicting the future development of the condition based on the derived formula.

Abstract: The system includes a film frame carrier (FFC) configured to support a workpiece, and the FFC is removably disposed on an FFC rotator. The system further includes an end effector configured to engage the FFC to remove the FFC from the FFC rotator and a scanner disposed in a movement path of the end effector. The scanner is configured to generate an overview image of the end effector engaged with the FFC as the end effector moves away from the FFC rotator. The system further includes a processor configured to receive the overview image from the scanner, determine an alignment between the FFC and the end effector according to the overview image, and control a movement of the end effector according to the alignment between the FFC and the end effector.

Abstract: A method includes receiving a plurality of paired input images, wherein the paired images includes a first set of images from a first modality and a second set of images from a second modality, outputting a list of bounding boxes and labels in response to running an image-based object detection model, mapping each bounding box to a region of interest that is corresponding to the bounding box and associated with the second set of images, cropping the region of interest from the first and second set of images to generate a cropped first and second set of images, sending the cropped first set of images to a first encoder and a cropped second set of images to a second encoder, wherein the first encoder is configured for the first modality and the second encoder is configured for the second modality, outputting object-level embeddings for both the cropped first and second set of images utilizing encoders, identifying a loss function associated with the images, and in response to when a threshold is met, outputting

Abstract: Disclosed in the present invention is a temporal information enhancement-based method for 3D medical image segmentation, belonging to the field of medical image segmentation. The method provides a circle transformer module for extraction and fusion of temporal information, and uses temporal input to improve the training effect of a deep learning model, thereby effectively eliminating interference of similar features and blurred images. In a training phase, an input sample is a temporal sequence, the model training effect is enhanced by extracting temporal information, and segmentation results before and after the combination of temporal information are both constrained, so that the model is no longer temporally dependent. In comparison with a training method in which a single sample is input, the present invention can improve the accuracy of an encoder-decoder structure-based segmentation model without costs.

Abstract: The present disclosure describes systems and methods to correct for perspective calibration variations of a variable thickness specimen with a single camera extensometer in a video extensometer system. In some examples, the systems and methods compensate for a change between a reference characteristic, such as a calibration plane, and an actual physical characteristic, such as a testing plane associated with a surface of a test specimen, during a testing operation. In some examples, a correction value is applied to an output (e.g., measured dimensions of the imaged test specimen) to compensate for the difference between the reference characteristic and the physical characteristic.

Abstract: The embodiments of the present disclosure relates to a method of room layout, an electronic device and a computer readable storage medium. The method includes: collecting a current frame RGB image, and acquiring a depth map corresponding to the current frame RGB image and pose information of a head mounted device; detecting and determining at least one object in the collected current frame RGB image; and associating the at least one object in the current frame RGB image, the depth image and the pose information with a room layout map corresponding to a previous frame RGB image, and generating a room layout map corresponding to the current frame RGB image for enabling a user to calibrate and create a rendering based on the room layout map corresponding to the current frame RGB image.

Abstract: Interactive based on said set steps of authentication methods for the recognition of a person. During authentication, a previously stored enrollment image is presented on a display to the person. A candidate person is instructed to present a reproduced image of the same scene and/or object to a camera while the person is holding the camera (mobile camera for example) unsupported in free space with respect to the scene or object. Alternatively, the user can hold the object unsupported in free space with respect the camera using the camera, a candidate image of the viewed scene or object is captured and presented with the previously stored enrollment image. The candidate person aligns the candidate image with the previously stored enrollment image. On alignment, the candidate image is verified as an authentic image of the person and the candidate person is authenticated as the person previously enrolled.

Abstract: A processor-implemented method for attention-based depth completion includes receiving, by an artificial neural network (ANN), an input. The input includes an image and a sparse depth measurement. The ANN extracts multi-scale visual features of the input. The ANN applies a self-attention mechanism to the multi-scale visual features to generate a set of attended multi-scale visual features. The ANN generates a dense depth map based on the set of attended multi-scale visual features.

Abstract: A person authentication support system (100) includes: a measurement condition acquisition unit (112) for acquiring a measurement condition of a depth in a measurement apparatus capable of generating image information indicating an image of a photographing region including a face of a target person, and depth information indicating a depth; a region setting unit (115) for setting a processing region being a region serving as a processing target of the depth information in the photographing region according to the measurement condition; a depth acquisition unit (116) for acquiring depth information including a depth in the processing region being measured in the measurement condition; and a collation determination unit (117) for determining whether image information of the target person is used in collation processing, based on depth information of the processing region.

Abstract: A process to identify a representative sub-sample of a rock sample includes the following steps: digitally acquiring at low resolution first 2D or 3D images of the rock sample, and by a recomposition of the first images along a direction representing a digital sample; subdividing with a volumetric subdivision the digital sample defining 3D digital blocks having a volume of substantially similar value; analysing each 3D digital block on the basis of key points or texture descriptors obtained elaborating said plurality of the first images on the basis of the properties or structures of the rock, identifying two or more homogeneous classes of 3D digital blocks, each homogeneous class defining a portion of digital sample having 3D digital blocks equivalent to each other as to the technical characteristics of the rock; and selecting at least one representative block for each homogeneous class identified, locating and extracting a sub-sample from the rock sample.

Abstract: Described herein is a system and a method for collaborative training of a machine learning model for pose estimation. The system includes a remote computing device in communication with a local computing device, wherein the remote computing device is configured to instruct the local computing device to receive a first image; and generate a first pose datum as a function of the first image using a local machine learning model. The remote computing device is configured to receive, from the local computing device, the first pose datum; generate a second pose datum as a function of the first pose datum and the first image using a remote machine learning model; and configure the local computing device to retrain the local machine learning model as a function of the second pose datum.

Abstract: An evaluation apparatus includes an obtaining unit configured to obtain an image of an evaluation region including a film forming region on which a film is formed by a film forming process, and a processor configured to process the image for the evaluation. The processor is configured to output a feature concerning an abnormality in the image in accordance with a learned model. The image and design information representing a geometrical feature of the film forming region are input to the learned model.

Abstract: An information processing device performs: acquiring a face similarity indicating a similarity between a face of a first person and a face of a second person; acquiring a voice similarity indicating a similarity between a voice of the first person and a voice of the second person; calculating an integrated similarity by integrating the face similarity and the voice similarity, and determining the integrated similarity as a final similarity when the face similarity falls within an integrated range including a threshold which is used to determine whether the first person and the second person are identical to each other, and calculating the face similarity as a final similarity when the face similarity is out of the integrated range; and outputting the final similarity.

Abstract: An encoder is trained together with a perception component based on a training set comprising unannotated sensor data sets and annotated sensor data sets in a sequence of multiple training steps.

Abstract: In various examples, three-dimensional (3D) object or feature detection and localization for autonomous and semi-autonomous systems and applications is described herein. Systems and methods are disclosed that use different types of sensors, such as an image sensor and a LIDAR sensor, to determine information associated with objects, such as traffic objects (e.g., traffic signs, traffic signals, traffic markings, etc.). To determine the information for an object, image data is processed to determine a bounding shape associated with the object. The bounding shape is then used to determine a 3D shape, such as a frustum, corresponding to the object. Additionally, points data generated using the LIDAR sensor, such as an occupancy map and/or a point cloud, is processed to identify a portion of the points associated with (e.g., located within) the 3D shape. This portion of the points may then be used to determine the information associated with the object.