Abstract: A method of monitoring objects capable of wireless communications is provided. The method includes detecting an activity performed by a user of a wireless communications device, acquiring information identifying an expected set of objects which are associated with the detected activity, determining whether at least one of the objects in the expected set is not in proximity of the wireless communications device, and if so, notifying the user that at least one of the objects in the expected set is not in proximity of the wireless communications device.

Abstract: Methods and systems for processing medical images. One method includes, in response to startup of an application using an algorithm, creating a server process supporting a programming language associated with the algorithm and loading a plurality of deep learning models used by the algorithm into a memory of the server process to create in-memory models. The method also includes processing a first set of one or more medical images with the server process using the algorithm and at least one model selected from the in-memory models, maintaining the in-memory models in the memory of the server process after processing the first set of one or more medical images, and, in response to a request to process a second set of one or more medical images, processing the second set of one or more medical images using the algorithm and at least one of the in-memory models.

Abstract: Technologies and techniques for vehicle perception. Contours are extracted from image data, and a first and second contour of a current image and next image are determined along a plurality of image row boundaries, wherein the contours includes a plurality of contour points linked to each other via respective edges. An image velocity vector is applied to project each of the plurality of first contour points to the next image, wherein at least some of the projected first contour points are not located on an image row boundary. The projected contour points are processed to determine projected edges forming a projected link between each projected contour point. The projected contour points are moved along the projected link to align each of the moved projected contour points to a respective image boundary. Each of the moved projected contour points are then linked to each of the second contour points.

Abstract: A motion encoder comprising a joint, the joint comprising a first part (304); and a second part (302) movably mounted with respect to the first part (304) to change the joint between a first joint condition and a second joint condition; a first image capture device (ICD) (310) comprising an array of light sensing elements arranged to generate images; a processing unit; and a pattern surface (312), the pattern surface (312) defining or being arranged to define a pattern (314) which moves relative to the first ICD (310) between a first pattern position and a second pattern position as the joint changes between the first joint condition and a second joint condition, the pattern (314) comprising pattern features, wherein the first ICD (310) has a field of view encompassing the first pattern position and the second pattern position such that the pattern features are visible to the first ICD (310) in both the first pattern position and the second pattern position; and the processing unit is communicatively coupled

Abstract: Embodiments of this disclosure provide a mouth shape correction model, and model training and application methods. The model includes a mouth feature extraction module, a key point extraction module, a first video module, a second video module, and a discriminator. The training method includes: based on a first original video and a second original video, extracting corresponding features by using various modules in the model to train the model; and when the model meets a convergence condition, completing the training to generate a target mouth shape correction model. The application method includes: inputting a video in which a mouth shape of a digital-human actor is to be corrected and corresponding audio into a mouth shape correction model, to obtain a video in which the mouth shape of the digital-human actor in the video is corrected, wherein the mouth shape correction model is a model trained by using the training method.

Abstract: A device and a method for determining a portion of broken grain and/or non-grain components in a stream of harvested material is disclosed. The device includes a camera to generate images of the stream of harvested material and an evaluation unit to estimate the portion in an image supplied by the camera. The evaluation unit includes a first-order classifier to estimate a first parameter of the stream of harvested material, and a plurality of second-order classifiers assigned to various values of the first parameter to estimate the portion in a stream of harvested material that has the assigned parameter value. The evaluation unit, with assistance of the first-order classifier, estimates a value of the first parameter using a first number of images of the camera, and then selects the second-order classifier assigned to the value of the first parameter to estimate the portion with assistance of the selected second-order classifier.

Abstract: A notification system notifies users when there is a match between images. The notification system includes a network interface receiving images from a plurality of user devices including a first user device and a second user device, a memory storing database and instructions, and a processor. The processor executes the instructions to function as an image processing unit processing the received images, a face recognition unit recognizing one or more faces in each received image, a face matching unit checking whether a face in a first image from the first user device matches a face in a second image from the second user device, a metadata processing unit processing metadata in each received image, and a notification unit notifying the first user device and the second user device in a case where the face in the first image matches the face in the second image.

Abstract: Described herein is a method for determining a filling level in at least one storage unit. The method includes illuminating the storage unit with at least one illumination pattern; selecting at least one first reflection feature and at least one second reflection feature; for the first reflection feature, generating at least two first sensor signals, and for the second reflection feature, generating at least two second sensor signals; evaluating the two first sensor signals and the two second sensor signals; determining at least one position of the first reflection feature and at least one position of the second reflection feature and determining at least one first vector of the first reflection feature and at least one second vector of the second reflection feature; and determining from the first vector and the second vector at least one elevation map and determining therefrom the filling level in the storage unit.

Abstract: A method to operate a wireless access point (WAP) that includes identifying multiple stations from uplink statistics of a plurality of uplinks and determining an optimal spatial state of an antenna array for reception of the plurality of uplinks. Channel state information (CSI) is evaluated for each antenna in the antenna array. Determining the optimal spatial state of the antenna array includes using the CSI for each antenna to extrapolate the optimal spatial state of the antenna array for the plurality of uplinks. The method further includes changing a spatial state of the antenna array to the optimal spatial state.

Abstract: An accommodation device with an air blowing function includes: an accommodation portion that is formed with an opening portion on an upper surface and accommodates a portable electronic device; a support portion, arranged so as to protrude from a side surface of the accommodation portion into the accommodation portion, the support portion configured to contact with side surfaces or upper and lower surfaces of the portable electronic device accommodated in the accommodation portion and hold the portable electronic device so as to pinch the portable electronic device; and a blower unit that blows air into the accommodation portion. The support portion has an air outlet for guiding the air sent out by the blower unit to the accommodation portion.

Abstract: A first image showing a non-frontal view is received. CNN outputs generated based on the first image are concatenated to generate a concatenated CNN output. The concatenated CNN output is input to a first encoder to generate a first value for a first statistical property and a first value for a second statistical property. Configural information is generated based on the first image. The configural information is input to a second encoder to generate a second value for the first statistical property and a second value for the second statistical property. Vectors are generated based on the first and second values for the first statistical property, and the first and second values for the second statistical property. A latent vector is generated based on the vectors. The latent vector is input to a decoder to generate a second image showing a frontal view.

Abstract: According to the present invention, there is provided an information processing system (100) including an acquisition unit (11) that acquires an image, which is generated by a camera, of a face, a selection unit (14) that extracts a person in the acquired image, and selects a person who performs a settlement process based on at least one of a distance from a checkout apparatus, an appearance frequency, and an appearance period of time in a case where a plurality of persons are extracted, an identification unit (12) that identifies the selected person based on the image in which the person is included, and database information in which a feature value of a face is registered, a settlement unit (checkout unit (21)) that performs the settlement process with respect to the identified person.

Abstract: Computerized methods are performable by a driver assistance system while the host vehicle is moving. The driver assistance system includes a camera connectible to a processor. First and second image frames are captured from the field of view of the camera. Corresponding image points of the road are tracked from the first image frame to the second image frame. Image motion between the corresponding image points of the road is processed to detect a hazard in the road. The corresponding image points are determined to be of a moving shadow cast on the road to avoid a false positive detection of a hazard in the road or the corresponding image points are determined not to be of a moving shadow cast on the road to verify detection of a hazard in the road.

Abstract: A method for locating a source of interference external to a wireless telecommunications network includes assigning first and second elevation values to each pixel in a grid of pixels that represents an area associated with a plurality of cellular antennas, and calculating at least one probability value that the source of interference external to the wireless telecommunications network is located at an associated pixel by calculating respective probability values for pixels in the grid using the first elevation value and the second elevation value of the associated pixel and vertical gain values for each of the plurality of cellular antennas.

Abstract: Processing photogrammetry data that represents parcels by computing a real-world distance between a first location representing a location of a specific separation point of a specific subarea included in the area as appearing in the initial representation and a second location that represents a location of the same specific separation point as appearing in the photogrammetry data, defining a specific separation point as identifying incorrect if the real-world distance exceeds a reference distance dictated by accuracy level, intersecting subareas of the subarea to which the examined vertex belongs, measuring a distance between the examined vertex and a closest point in the intersecting subarea, marking the examined vertex as an error if the intersecting distance exceeds a threshold, computing angles for edges defining subarea boundaries, and removing a vertex between two subsequent edges having an angle between them that satisfies a condition representing that the subsequent edges constitute a single line.

Abstract: Techniques for improving a blast pattern at a mining site include conducting an initial blast and recording the initial blast as a high speed optical video. The high speed optical video, and the blast pattern used in the initial blast are sent as inputs to a machine learning model, which correlates one or more characteristics of the region being blasted with measurements associated with characteristics of the region being blasted obtained from the high speed optical video. The machine learning model can then determine an improved blast pattern based on the correlation made. This improved blast pattern can be displayed on a user computing device, or transmitted to a drilling system to automatically drill the improved blast pattern for subsequent blasts.

Abstract: An in-vehicle entertainment system having functionality of operation interface switchable is disclosed. The system comprises an in-vehicle host electronic device and a hub having a host-to-host transmission function. After at least two mobile electronic devices are coupled to the hub, the in-vehicle host electronic device firstly prioritizes a communication order between the two mobile electronic devices, and then lets a display unit show an operation interface (like CarPlay operation interface) corresponding to the mobile electronic device that is prioritized with a first ranked communication order. Moreover, in case of a peripheral electronic device being coupled to the hub as well as the display unit being showing the operation interface, the in-vehicle host electronic device is allowed to access a peripheral electronic device through the hub.

Abstract: A notification system notifies users when there is a match between images. The notification system includes a network interface receiving images from a plurality of user devices including a first user device and a second user device, a memory storing database and instructions, and a processor. The processor executes the instructions to function as an image processing unit processing the received images, a face recognition unit recognizing one or more faces in each received image, a face matching unit checking whether a face in a first image from the first user device matches a face in a second image from the second user device, a metadata processing unit processing metadata in each received image, and a notification unit notifying the first user device and the second user device in a case where the face in the first image matches the face in the second image.

Abstract: A cell image analysis method may include: obtaining, for each of cell images, a value of a feature parameter to be used in determination of a type of a cell, by analyzing the cell images; and displaying the value of the feature parameter in association with the each of the cell images.

Abstract: A compiler manages memory usage in the machine learning accelerator by intelligently ordering computations of a machine learning network. The compiler identifies partial networks of the machine learning network representing portions of the machine learning network across multiple layers on which an output or set of outputs are dependent. Because any given output may depend on only a limited subset of intermediate outputs from the prior layers, each partial network may include only a small fraction of the intermediate outputs from each layer. Instead of implementing the MLN by computing one layer at a time, the compiler schedules instructions to sequentially implement partial networks. As each layer of a partial network is completed, the intermediate outputs can be released from memory. The described technique enables intermediate outputs to be directly streamed between processing elements of the machine learning accelerator without requiring large transfers to and from external memory.