Abstract: Systems, devices, and methods for training a neural network and performing image interest point detection and description using the neural network. The neural network may include an interest point detector subnetwork and a descriptor subnetwork. An optical device may include at least one camera for capturing a first image and a second image. A first set of interest points and a first descriptor may be calculated using the neural network based on the first image, and a second set of interest points and a second descriptor may be calculated using the neural network based on the second image. A homography between the first image and the second image may be determined based on the first and second sets of interest points and the first and second descriptors. The optical device may adjust virtual image light being projected onto an eyepiece based on the homography.

Abstract: A deep neural network(s) (DNN) may be used to detect objects from sensor data of a three dimensional (3D) environment. For example, a multi-view perception DNN may include multiple constituent DNNs or stages chained together that sequentially process different views of the 3D environment. An example DNN may include a first stage that performs class segmentation in a first view (e.g., perspective view) and a second stage that performs class segmentation and/or regresses instance geometry in a second view (e.g., top-down). The DNN outputs may be processed to generate 2D and/or 3D bounding boxes and class labels for detected objects in the 3D environment. As such, the techniques described herein may be used to detect and classify animate objects and/or parts of an environment, and these detections and classifications may be provided to an autonomous vehicle drive stack to enable safe planning and control of the autonomous vehicle.

Abstract: The present invention relates to a scene descriptor generation system, including an image sensor including a plurality of pixels, successively supplying S binary image planes, each including a single binary value per pixel; and a processing circuit configured to, for each binary image plane supplied by the sensor (100), to implement the successive steps of: a) calculating N convolutions of the binary image plane with respectively N distinct convolution kernels, to provide N convolved images; b) generating at least one meta-image from the N convolved images supplied at step a); and c) for each of the N convolved images supplied at step a), generating a binary descriptor from said convolved image and from said at least one meta-image.

Abstract: A position where a user is located can be estimated easily and quickly. When a hold detection unit 122 detects that the user holds a mobile terminal 100 by hand to use the mobile terminal 100 that is held by at least one hand of the user when used by the user and has a camera 130 on the surface facing the ceiling when used by the user, the estimation unit 240 estimates a position where a ceiling is located as the position where the user is located by using an image of the ceiling captured by the camera 130.

Abstract: A medical information processing apparatus includes processing circuitry. The processing circuitry is configured to acquire medical image data representing a blood vessel of a subject, extract a blood vessel shape from the medical image data, determine degree of meandering in each of regions from the extracted blood vessel shape, specify a deformed region in which the degree of meandering changes due to insertion of a device into the blood vessel on the basis of the degree of meandering, and output the deformed region in the blood vessel.

Abstract: An information processing apparatus (2000) detects an estimation position (24), which is a position where a target object is estimated to be present, from a captured image (20) generated by a camera (10). The target object is a belonging of a person. The information processing apparatus (2000) detects a plurality of parts (22) constituting a body of a person included in the captured image (20). For example, the parts (22) are joints. The information processing apparatus (2000) detects the estimation position (24) of the target object in the captured image (20) using information relating to a position of each of a plurality of the detected parts (22).

Abstract: Various systems and methods for for clustering facial images in, for example, surveillance systems.

Abstract: The present disclosure relates to methods and systems to generate a modified feature map specific to a client. A template feature map may be modified based on usage data associated with a client. The template feature map may represent a visual representation of a plurality of features provided by an operator, each feature associated with a plurality of instructions to be processed for the client. The usage data may be compared with each feature to determine whether any feature is associated and/or utilized by the client. Based on determining whether the usage data indicates that any feature is associated and/or utilized by the client, the template feature map may be modified to perform an action to the template feature map indicating that a feature is associated and/or utilized by the client. A modified template feature map may be generated that is specific for a client.

Abstract: Systems and methods for provisioning real-time three-dimensional maps, including: at a first processing stage: updating a three-dimensional map of an environment based on received point cloud data; extracting three-dimensional proposals of objects of the environment; projecting the three-dimensional proposals onto a two-dimensional image of the environment; generating two-dimensional proposals of the objects of the environment; at a second processing stage: detecting the objects of the environment from the two-dimensional image of the environment; generating two-dimensional bounding boxes of the objects of the environment; matching the two-dimensional bounding box with a particular two-dimensional proposal of the two-dimensional proposals; and labeling, for each two-dimensional proposal that is matched to a two-dimensional bounding box, the three-dimensional proposal corresponding to the two-dimensional proposal with semantic information associated with the two-dimensional bounding box that is matched to the

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for processing data generated by a sensing system that rotationally senses an environment. In one aspect, a method comprises partitioning a predetermined period of time into a plurality of sub-periods, wherein the predetermined period of time is a period of time for which data generated by the sensing system constitutes a complete rotational sensing of the environment; for each sub-period: receiving current data generated by the sensing system during the sub-period and characterizing a respective partial scene of the environment; processing the current data using an object detection neural network to generate a current object detection output that is specific to the respective partial scene of the environment.

Abstract: An information processing apparatus (1) includes: a presence information acquisition unit (2) that acquires presence information that is generated based on images shot for each of a plurality of predetermined places and indicates a place where a person is present at a point in time of the shooting of the images; and a presence place estimation unit (3) that estimates a place where a specific person is currently present from among the plurality of predetermined places based on the acquired presence information. The presence information includes time information at the point in time of the shooting. The presence place estimation unit (3) estimates the place where the specific person is currently present using a presence probability of the specific person at the predetermined places, the presence probability being calculated based on an amount of time elapsed from the point in time of the shooting.

Abstract: A video surveillance system and computer assisted method stores video streams of numerous video cameras installed throughout a facility (e.g., a school, an airport, etc.) and batch searches those video streams for relevant video frames that may show an individual with certain predetermined facial features. In some examples, the search range is limited by a specified time window, by certain designated cameras, and/or by certain predetermined areas of the facility. From those video streams, a collection of video clips containing the relevant video frames are identified and batch exported for later analysis by video surveillance personnel. In some examples, the collected video clips are compiled into a single composite video with the video clips being arranged by time for convenient and efficient viewing. In addition or alternatively, in some examples, the video clips are stored as separate files in a directory for viewing the video clips individually.

Abstract: The present disclosure is directed to a method and system for automatically predicting a physiological parameter based on images of vessel. The method includes receiving the images of a vessel acquired by an imaging device. The method further includes determining a sequence of temporal features at a sequence of positions on a centerline of the vessel based on the images of the vessel, and determining a sequence of structure-related features at the sequence of positions on the centerline of the vessel. The method also includes fusing the sequence of structure-related features and the sequence of temporal features at the sequence of positions respectively. The method additionally includes determining the physiological parameter for the vessel at the sequence of positions, by using a sequence-to-sequence neural network configured to capture sequential dependencies among the sequence of fused features.

Abstract: The disclosure pertains to techniques for image processing. One such technique comprises a method for image selection, comprising obtaining a sequence of images, detecting one or more objects in one or more images of the sequence of images, determining a location for each detected object in the one or more images, determining a trajectory for each detected object based on a determined location for each respective detected object in two or more images of the sequence of images, determining a trajectory waypoint score for the trajectory of each detected object, determining a set of selected images based on an aggregation of trajectory waypoint scores for each detected object in each respective image, and outputting the set of selected images for presentation.

Abstract: A method for 3D auto-labeling of objects with predetermined structural and physical constraints includes identifying initial object-seeds for all frames from a given frame sequence of a scene. The method also includes refining each of the initial object-seeds over the 2D/3D data, while complying with the predetermined structural and physical constraints to auto-label 3D object vehicles within the scene. The method further includes linking the auto-label 3D object vehicles over time into trajectories while respecting the predetermined structural and physical constraints.

Abstract: In part, the disclosure relates to an automated method of branch detection with regard to a blood vessel imaged using an intravascular modality such as OCT, IVUS, or other imaging modalities. In one embodiment, a representation of A-lines and frames generated using an intravascular imaging system is used to identify candidate branches of a blood vessel. One or more operators such as filters can be applied to remove false positives associated with other detections.

Abstract: A circuit, system, and method for measuring or detecting pressure or force of a fingerprint on an array of electrodes is described. Pressure or force may be measured or detected using a processed image of the fingerprint, or by measurement of capacitance of deformed variable capacitors.

Abstract: An image processing device has: a change detecting unit configured to, based on a captured image in which an product to be monitored is shown, detect a change in display status of the product; and a display detecting unit configured to detect that the product is returned to a different place from a place where the product has been picked up based on the change in display status of the product detected by the change detecting unit and on a person included in the captured image.

Abstract: There is provided an apparatus (100) for determining a location of a personal care device on a user. The apparatus (100) comprises a processor (102) configured to detect a location of one or more parts of the body of a user, detect motions of at least one of the parts of the body of the user induced by the personal care device when the personal care device is in use and determine the location of the personal care device on the user based on the detected motions of the at least one part of the body of the user in relation to the detected location of the at least one part of the body of the user. The processor (102) is configured to acquire images of the user from an imaging sensor (108) directed at the user, and detect the location of the one or more parts of the body of a user and the motions of the at least one part of the body of the user induced by the personal care device from the acquired images of the user.

Abstract: A method for improving geometry-based monocular structure from motion (SfM) by exploiting depth maps predicted by convolutional neural networks (CNNs) is presented. The method includes capturing a sequence of RGB images from an unlabeled monocular video stream obtained by a monocular camera, feeding the RGB images into a depth estimation/refinement module, outputting depth maps, feeding the depth maps and the RGB images to a pose estimation/refinement module, the depths maps and the RGB images collectively defining pseudo RGB-D images, outputting camera poses and point clouds, and constructing a 3D map of a surrounding environment displayed on a visualization device.