Abstract: Various embodiments related to an electronic device are disclosed, and an electronic device according to one embodiment comprises: a camera; at least one microphone; at least one motor for rotating the camera around at least one axis; and at least one processor, wherein the processor can be configured to: receive a voice through at least one microphone and an image through the camera; perform a first control on at least one motor on the basis of at least a first motor control mode; use at least one motor on the basis of at least the first control so as to perform first face tracking using the received image; perform a second control on at least one motor on the basis of a second motor control mode during the first tracking; and perform voice recognition or face recognition by using at least one motor on the basis of at least the second control. Other various embodiments are possible.

Abstract: A method for identifying body representation information in an image, includes: determining a skeleton-like line of a body in the image; and then performing identification of the body representation information according to the skeleton-like line. An apparatus, and a device for identifying body representation information in an image and a storage medium are also provided.

Abstract: Embodiments of the present disclosure disclose an image processing method and apparatus, and an electronic device. The method includes: obtaining a key frame image that includes a target object, and obtaining a to-be-processed frame image that includes the target object; extracting a feature point in the key frame image and a feature point in the to-be-processed frame image, respectively; determining a matching relationship between each feature point in the key frame image and each feature point in the to-be-processed frame image; determining a transformation relationship based on the feature point in the key frame image and a feature point in the to-be-processed frame image that matches the feature point in the key frame image; and processing the to-be-processed frame image based on the transformation relationship to obtain a target frame image.

Abstract: Methods, systems and apparatus, including computer programs encoded on computer storage media for determining asset efficiency. Unmanned Aerial Vehicles (UAVs) may be used to obtain aerial images of locations, property or structures. The aerial images may be geo-rectified, and a ortho-mosaic, digital surface model, or a point cloud may be created. In the context of an operation where mobile assets are used, such as construction or earth moving equipment, location-based event information may be obtained. The system determines efficiency clusters for particular assets, and provides an exploration interface to present and navigate via the efficiency cluster.

Abstract: Techniques for dialog data collection are disclosed. In an embodiment, a method comprises providing a first graphical user interface (10) configured to receive first user input data, providing a second graphical user interface (20) configured to receive second user input data, asynchronously transmitting the first user input data to the second graphical user interface (20) or the second user input data to the first graphical user interface (10), and generating training data for a natural language processing system model (60) based on the first user input data and the second user input data.

Abstract: In some examples, image content moderation may include classifying, based on a learning model, an object displayed in an image into a category. Further, image content moderation may include detecting, based on another learning model, the object, refining the detected object based on a label, and determining, based on the another learning model, a category for the refined detected object. Further, image content moderation may include identifying, based on the label, a keyword associated with the object, and determining, based on the identified keyword, a category for the object. Further, image content moderation may include categorizing, based on a set of rules, the object into a category, and moderating image content by categorizing, based on aforementioned analysis the object into a category. Yet further, image content moderation may include tagging, based on fusion-based tagging, the object with a category and a color associated with the object.

Abstract: A system for machine learning that is configured to receive an input having a plurality of features and predict one or more attributes of the input. The system includes a security mechanism, which determines an initial value for each of the features; determines a perturbation value for each of the features, the perturbation being randomly selected; adds the perturbation value to the initial value to determine a perturbed value for each of the features; and quantizes the perturbation value for each of the features to determine a quantized value for each of the features. The system also includes a classifier that receives the quantized value for each of the features and predict the one or more attributes of the input based on the quantized value for each of the features.

Abstract: A device may receive document image data that includes an image of a document to be digitized. The device may detect, from the document image data, a table of information that is depicted in the image. The device may determine a data extraction score associated with a table image, wherein the data extraction score is associated with using a data conversion technique to convert the table image to digitized table data. The device may perform, based on the data extraction score not satisfying a threshold, a morphological operation on the table image to generate an enhanced table image that corresponds to an enhanced table of information associated with the table of information. The device may process, using the data conversion technique, the enhanced table image to extract the information from the enhanced table. The device may perform an action associated with the extracted information.

Abstract: There is provided methods and systems for controlling airplane taxiing, including detecting aerodrome taxiway lines and estimating cross-track error.

Abstract: The present invention concerns a method for the 3D reconstruction of a scene comprising the matching (610) of a first event from among the first asynchronous successive events of a first sensor with a second event from among the second asynchronous successive events of a second sensor depending on a minimisation (609) of a cost function (E).

Abstract: An information processing apparatus includes a processor configured to: acquire a read image and item information, the read image being an image obtained by reading a paper medium including plural documents, the item information being information of an item specified by a user from among plural items contained in the documents; extract a character string from the read image, the character string being associated with the item information; if a character string contained in a page of the read image and extracted from the page differs from a character string extracted from the previous page immediately preceding the page, set a split position, the split position being a position at which to split out a portion of the read image as a set of documents, the portion being a portion of the read image from a page where extraction has begun to the previous page; and output the read image split in accordance with the split position.

Abstract: The present disclosure relates to phenotype analysis of cellular image data using a deep metric network. One example embodiment includes a method. The method includes receiving a target image of a target biological cell having a target phenotype. The method also includes obtaining a semantic embedding associated with the target image. The semantic embedding is generated using a machine-learned, deep metric network model. Further, the method includes obtaining, for each of a plurality of candidate images of candidate biological cells each having a respective candidate phenotype, a semantic embedding associated with the respective candidate image. In addition, the method includes identifying, for each of the semantic embeddings, common morphological variations and reducing, for each of the semantic embeddings based on the identified common morphological variations, effects of nuisances. Even further, the method includes determining, by the computing device, a similarity score for each candidate image.

Abstract: A computer-implemented method of detecting logos in a graphical rendering may comprise detecting, using a first and a second trained object detector, logos in the graphical rendering and outputting a first and a second list of detections and filtering, using at least a first and a second prior performance-based filter, the received first and second lists of detections into a first group of kept detections, a second group of discarded detections and a third group of detections. Detections in the third group of detections may be clustered in at least one cluster comprising detections that are of a same class and that are generally co-located within the electronic image. A cluster score may then be assigned to each cluster. A set of detections of logos in the graphical rendering may then be output, the set comprising the detections in the first group and a detection from each of the clusters whose assigned cluster score is greater than a respective threshold.

Abstract: Embodiments are provided that include receiving sensor data from a sensor positioned at a plurality of positions in an environment. The environment includes a plurality of landmarks. The embodiments also include determining, based on the sensor data, a subset of the plurality of landmarks detected at each of the plurality of positions. The embodiments further include determining, based on the subset of the plurality of landmarks detected at each of the plurality of positions, a detection frequency of each landmark. The embodiments additionally include determining, based on the determined detection frequency of each landmark, a localization viability metric associated with each landmark. The embodiments still further include providing for display, via a user interface, a map of the environment. The map includes an indication of the localization viability metric associated with each landmark.

Abstract: An image processing apparatus includes a search unit that searches a search area, which is a part of a first image extending over two pages included in the first image, for a dividable area, the first image being indicated by first image data and a division unit that generates second image data indicating two second images, each of which corresponds to one of the two pages, by dividing the first image along a dividing line passing through, in a second direction, which is perpendicular to a first direction, the dividable area found by the search unit, the first direction being a direction across the two pages.

Abstract: The present disclosure provides a method, a device, an apparatus and storage medium for facial matching, wherein the method includes: acquiring an image to be matched; conducting matching for the image to be matched based on at least one of an original sample database and an associative sample database; and outputting a final matching result, wherein the original sample database includes an original sample image, and the associative sample database includes an associative sample image which is formed by adding an associative feature to the original sample image. Herein, obtaining the original sample database and the associative sample database comprises: acquiring the original sample image; obtaining the original sample database based on the original sample image; adding the associative feature to the original sample image in the original sample database and generating the associative sample image, to obtain the associative sample database.

Abstract: Systems, devices, media and methods are presented for a human pose tracking framework. The human pose tracking framework may identify a message with video frames, generate, using a composite convolutional neural network, joint data representing joint locations of a human depicted in the video frames, the generating of the joint data by the composite convolutional neural network done by a deep convolutional neural network operating on one portion of the video frames, a shallow convolutional neural network operating on a another portion of the video frames, and tracking the joint locations using a one-shot learner neural network that is trained to track the joint locations based on a concatenation of feature maps and a convolutional pose machine. The human pose tracking framework may store, the joint locations, and cause presentation of a rendition of the joint locations on a user interface of a client device.

Abstract: A live facial recognition method includes capturing a zoom-out image of a face of a subject under recognition; and detecting a frame outside the face of the subject under recognition on the zoom-out image. The subject under recognition is determined to be a living subject when the zoom-out image includes no frame outside the face.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing images or features of images using an image classification system that includes a batch normalization layer. One of the systems includes a convolutional neural network configured to receive an input comprising an image or image features of the image and to generate a network output that includes respective scores for each object category in a set of object categories, the score for each object category representing a likelihood that that the image contains an image of an object belonging to the category, and the convolutional neural network comprising: a plurality of neural network layers, the plurality of neural network layers comprising a first convolutional neural network layer and a second neural network layer; and a batch normalization layer between the first convolutional neural network layer and the second neural network layer.

Abstract: In one embodiment, a system to accelerate batch-normalized convolutional neural network (CNN) models is disclosed. The system extracts a plurality of first groups of layers from a first CNN model, each group of the first groups having a first convolutional layer and a first batch-norm layer. For each group of the plurality of first groups, the system calculates a first scale vector and a first shift vector based on the first batch-norm layer, and generates a second convolutional layer representing the corresponding group of the plurality of first groups based on the first convolutional layer and the first scale and the first shift vectors. The system generates an accelerated CNN model based on the second convolutional layer corresponding to the plurality of the first groups, such that the accelerated CNN model is utilized subsequently to classify an object perceived by an autonomous driving vehicle (ADV).

Abstract: Images are intelligently selected to create image narratives. Instead of a user having to manually search and locate images to view, the images to associate with a particular image narrative are programmatically determined. Many different types of image narratives may be created. For example, one image narrative may show images that include both a first user and a second user over some period of time. Another image narrative may show images that relate to an activity that a first user enjoys or an event that included the user (e.g., a graduation). The tags and metadata associated with the images of the user are analyzed to determine the tags that are important to the user. For example, the importance might be determined based on the frequency of the tags within the images. After creation, the user may select one of the image narratives to view the associated images.

Abstract: Provided is an information processing system, including a processor, a storage device and a display device. The storage device stores images captured by imaging devices. If an image of a moving object is selected as a search query, then the processor executes a similarity search on the images stored in the storage device, with the selected image as the search query. The display device displays images attained as results of the similarity search. If an image of a moving object is added as a search query, then the processor executes a similarity search with all selected images as search queries. The display device updates images to be displayed such that images attained as results of the similarity search performed with all the selected images as the search queries are displayed.

Abstract: A comparator may automatically detect and quantify subtle and/or microscopic variance to a feature of a three-dimensional (“3D”) object in a reproducible manner based on point cloud imaging of that 3D object. The comparator may isolate a first set of data points, that represent the object feature at a first time, in a reference point cloud, and may isolate a second set of data points, that represent the same but altered object feature at a different second time, in a non-reference point cloud. The comparator may detect variance between positional values and visual characteristic values of the second set of data points and the corresponding positional values and visual characteristic values of the first set of data points, and may quantify a change occurring to the object feature between the first time and the second time based on a mapping of the variance to a particular unit of measure.

Abstract: In one aspect, the present disclosure provides a method for labeling one or more medications concurrently administered to a patient as a line of therapy. The method includes identifying medical records of the patient from a plurality of digital records, creating, from the subset of medical records, a plurality of treatment intervals including at least one medication administered to the patient and a time interval, associating medications of the one or more treatments with a respective treatment interval when the administration of the medication falls within the time interval, refining the time interval of a respective treatment interval when a treatment of the one or more treatments falls outside the time interval but within an extension period, identifying one or more potential lines of therapy from the plurality of treatment intervals, and labeling the potential line of therapy having the highest maximum likelihood estimation as the line of therapy.

Abstract: An image adaptive feature extraction method includes dividing an image into a plurality of blocks, performing a feature extraction processing on the plurality of blocks, and obtaining a block feature from each of the plurality of blocks after the feature extraction processing; calculating each block feature by means of a support vector machine (SVM) classifier, wherein each block feature is calculated to obtain a hyperplane normal vector; setting a threshold value, determining the block feature according to the hyperplane normal vector, recording the block as an adaptive feature block when a value of the hyperplane normal vector is higher than the threshold value, and integrating each adaptive feature block to form an adaptive feature image. Because an image adaptive feature extraction process is performed before a pedestrian image detection is calculated, and effective feature data is then selected, computational efficiency is boosted and detection pedestrian error probability is reduced.

Abstract: In one aspect, a retail store has multiple sensors, including item sensors in a shopping cart for gathering data from a shopper-selected first item. At least certain of the sensor data is provided to a classifier, which was previously-trained (using data including optical data from known items) to identify possible item matches corresponding to data sensed from the first item. An item identification hypothesis that the shopper-selected first item has a particular identity is evaluated based on (a) information from the classifier, and (b) store layout data indicating items associated with a store location visited by the cart or shopper. The item identification hypothesis has a confidence score. If the score meets a criterion, an item of the hypothesized identity is added to a shopping tally. A great number of other features and arrangements are also detailed.

Abstract: A system and process to determine whether a digital image has been manipulated involves determining expected locations of non-standard pixels in the digital image, and determining a feature for evaluating the non-standard pixels. The feature in pixels of the digital image that are located at the expected locations of non-standard pixels is then measured, and a statistical measure of the feature of pixels in the digital image that are located at the expected locations of non-standard pixels is evaluated. The digital image is assessed to determine a probability that the digital image includes a manipulated portion, based on the statistical measure. The system and process can also determine a make and model of an image sensing device via an examination of the non-standard pixels.

Abstract: The present disclosure provides a method and device for acquiring 3D information of an object. The method includes: extracting two-dimensional (2D) key points of the object based on the image, and determining a candidate 3D model set matching the image; determining a plurality of first reference attitudes and positions of each candidate 3D model according to the 3D key points and the 2D key points; acquiring a plurality of reprojection error values between each candidate 3D model and the object at the plurality of first reference attitudes and positions; determining a first target attitude and position and a first target 3D model corresponding to a minimum reprojection error value in the first reprojection error value set; and acquiring the 3D information of the object based on the first target attitude and position and the first target 3D model.

Abstract: A display apparatus and method that reduces block noise by performing block noise boundary detections and generating a block noise boundary map on the basis of a result of the detections to cope with local random block noise having irregular shaped and blurred block boundaries to perform adaptive deblocking filtering. The apparatus includes: an image receiver; a map generator to generate a block boundary map by performing convolution using a plurality of kernels on a received image; a determiner to determine a filter parameter on the basis of the block boundary map and a block boundary period included in the block boundary map; a deblocking filter to vary a filter strength on the basis of the determined filter parameter; and a display on which an image in which block noise is removed by the deblocking filter is displayed.

Abstract: The present disclosure provides an expression recognition method, a computer device, and a computer-readable storage medium. The expression recognition method includes: obtaining a facial image to be recognized; preprocessing the facial image to be recognized to obtain a preprocessed facial image; obtaining a first output image according to the preprocessed facial image, wherein the first output image at least represents a texture feature of the facial image to be recognized; obtaining a second output image according to the first output image, wherein the second output image at least represents a structural feature of organs of the facial image to be recognized; and determining an expression category corresponding to the facial image to be recognized, according to the second output image.

Abstract: A processing device includes a processor core and a number of calculation modules that each is configurable to perform any one of operations for a convolutional neuron network system. A first set of the calculation modules are configured to perform convolution operations, a second set of the calculation modules are reconfigured to perform averaging operations, and a third set of the calculation modules are reconfigured to perform dot product operations.

Abstract: A state of a sample surface is accurately determined without lowering analysis efficiency. There is provided an apparatus for determining a state of a sample to be analyzed contained in a container, in which the apparatus acquires an image of the sample, analyzes a position and a size of an object to be detected with respect to a detection range set in the image by using the image of the sample, and determines the state of the sample based on a result of the analysis.

Abstract: A widget recommendation system and method recommends user interface widgets for an event that contains content or a presentation. In one embodiment, the system recommends user interface widgets for a landing page for the content or presentation of an event. The system and method may extract features from past events and recommend the user interface widgets.

Abstract: A method for detection of an emitter mobile device (B) equipped with a light by a receiver mobile device (A) equipped with a camera, comprising: the emitter mobile device (B) emits a light-modulated signal to advertise its identity or state; the receiver mobile device (A) captures a series of frames with said camera; the modulated signal is detected from said frames; the pixel position of the emitter mobile device on a within the camera frame of the receiver mobile device is detected.

Abstract: Methods are devices are provided for setting an identity image of an account. In the method, an identity image setting operation for a target account is detected; an identity image set is displayed in response to detection of the identity image setting operation, the identity image set including at least one identity image to be selected; selection operation over the identity image to be selected is detected; and an identity image of the target account is set in response to detection of the selection operation according to a target image indicated by the selection operation.

Abstract: An image processing system may include an imaging device for capturing an image and an image processing apparatus for processing the image. The imaging device may include an imaging unit for capturing the image, a first recording unit for recording information relating to the image, the information being associated with the image, and a first transmission control unit for controlling transmission of the image to the image processing apparatus. The image processing apparatus may include a reception control unit for controlling reception of the image transmitted from the imaging device, a feature extracting unit for extracting a feature of the received image, a second recording unit for recording the feature, extracted from the image, the feature being associated with the image, and a second transmission control unit for controlling transmission of the feature to the imaging device.

Abstract: A container crane has a trolley movable on a cross-member of a gantry. A harness for picking up and setting down a container and at least one laser scanner are arranged on the trolley. The laser scanner captures a depth image which, as a function of a first and a second angle, indicates the distance of object points detected by a laser beam. The captured depth image is evaluated. Based on the object points, objects are detected and their locations are determined. The objects comprise the harness and/or a container picked up by the harness, and further objects. Based on the detected object points, the contour of the harness and/or of the container picked up by the harness is determined. Detection of further objects within regions defined by the contour is suppressed. A control device takes the detected objects and their locations into account for controlling the container crane.

Abstract: A system and method for geolocation-based pictographs are provided. In example embodiments, a current geolocation of a user device is determined. A pictograph is identified based on the current geolocation of the user device. The identified pictograph is presented on a user interface of the user device.

Abstract: An imaging apparatus includes an imaging device, a first imaging optical system and a second imaging optical system that form respective input images from mutually different viewpoints onto the imaging device, and a first modulation mask and a second modulation mask that modulate the input images formed by the first imaging optical system and the second imaging optical system. The imaging device captures a superposed image composed of the two input images that have been formed by the first imaging optical system and the second imaging optical system, modulated by the first modulation mask and the second modulation mask, and optically superposed on each other, and the first modulation mask and the second modulation mask have mutually different optical transmittance distribution characteristics.

Abstract: Region detection device includes acquirer that acquires a fingertip image of a fingertip including a nail, converter that transforms the fingertip image into a polar coordinate image in which one point inside the nail in the fingertip image is used as a pole, and detector that detects a region of the nail using the polar coordinate image.

Abstract: Systems, methods, and devices for verifying a user are disclosed. A speech-controlled device captures a spoken command, and sends audio data corresponding thereto to a server. The server performs ASR on the audio data to determine ASR confidence data. The server, in parallel, performs user verification on the audio data to determine user verification confidence data. The server may modify the user verification confidence data using the ASR confidence data. In addition or alternatively, the server may modify the user verification confidence data using at least one of a location of the speech-controlled device within a building, a type of the speech-controlled device, or a geographic location of the speech-controlled device.

Abstract: Aspects of the present disclosure are for an interaction mode system that provides multiple interaction modes in an artificial reality environment with automatic, context-specific transitions between interaction modes. An interaction mode can specify how the interaction mode system determines direction indications and movement within an artificial reality environment and interactions for making selections or performing other actions. In some implementations, the interaction mode system can control at least four interaction modes including a no-hands, 3DoF mode, a no-hands, 6DoF mode, a gaze and gesture mode, and a ray casting mode. The interaction mode system can employ a mapping of interaction mode context factors (e.g., which components are enabled, mode settings, lighting or other environment conditions, current body positions, etc.) to interaction modes to control transitioning between particular interaction modes.

Abstract: This disclosure is generally directed to identifying electronic forms using spatial information of elements presented on a website. Identifying a type of an electronic form may include identifying particular input elements associated with the form, determining a bounding region of the input element, expanding the bounding region, and determining any intersections of the expanded bounding region with one or more label elements proximate to the input element. Keywords of the label elements can be analyzed to increase or decrease a confidence level that an input element is associated with a particular input type. A bounding region can be dynamically sized based on a number of intersecting elements. An electronic form can be identified based on the identified input elements. In some instances, the electronic forms may assist a user in accessing or updating remotely stored personal information, including payment information, across one or more third party electronic sites.

Abstract: Methods and apparatus consistent with the invention provide the ability to organize, index, search, and present time series data based on searches. Time series data are sequences of time stamped records occurring in one or more usually continuous streams, representing some type of activity. In one embodiment, time series data is organized into discrete events with normalized time stamps and the events are indexed by time and keyword. A search is received and relevant event information is retrieved based in whole or in part on the time indexing mechanism, keyword indexing mechanism, or statistical indices calculated at the time of the search.

Abstract: Systems, methods and non-transitory computer-readable media for triggering actions based on touch-free gesture detection are disclosed. The disclosed systems may include at least one processor. A processor may be configured to receive image information from an image sensor, detect in the image information a gesture performed by a user, detect a location of the gesture in the image information, access information associated with at least one control boundary, the control boundary relating to a physical dimension of a device in a field of view of the user, or a physical dimension of a body of the user as perceived by the image sensor, and cause an action associated with the detected gesture, the detected gesture location, and a relationship between the detected gesture location and the control boundary.

Abstract: A processing apparatus configured to obtain polarization information using a plurality of images each of which is obtained in a different polarization state of light from an object includes a processor configured to select one reference image from the plurality of images and to generate difference data relating to a difference between the reference image and an image other than the reference image, and a recorder configured to record the reference image and the difference data.

Abstract: An information processing apparatus includes a receiving unit that receives an input picture obtained by capturing a landscape including an object, and an estimation unit that estimates an image represented by the landscape appearing in the input picture based on a learning model in which the input picture received by the receiving unit is input.

Abstract: A system and method for localizing an entity includes a processing system with at least one processing device. The processing system is configured to obtain sensor data from a sensor system that includes at least a first set of sensors and a second set of sensors. The processing system is configured to produce a map image with a map region that is selected and aligned based on a localization estimate. The localization estimate is based on sensor data from the first set of sensors. The processing system is configured to extract sets of localization features from the sensor data of the second set of sensors. The processing system is configured to generate visualization images in which each visualization image includes a respective set of localization features. The processing system is configured to generate localization output data for the vehicle in real-time by optimizing an image registration of the map image relative to the visualization images.

Abstract: A method for detecting defects in a product (10), such as food packaging, having a range of thicknesses of cross-section through which detection will take place; the method comprising: scanning the product (10), with for instance x-rays, to identify one or more light regions, and one or more dense regions of the product; and creating a first signal path and a second signal path from a single set of scanning data, and conditioning: the first signal path for detection of defects in the one or more dense regions; and the second signal path for detection of defects (14a, 14b, 14c) in the one or more light regions. Advantageously, detection of defects in the contents (dense region) of a product and the seal (light region) of the product is conducted simultaneously.

Abstract: A neuromorphic device is disclosed. The neuromorphic device may include an input element; a synapse element having a plurality of synapse blocks; a logic element; and an output element. The plurality of synapse blocks may share the logic element.