Abstract: An electronic device obtains video content and a textual query associated with a video moment in the video content. The video content is divided video segments, and the textual query includes one or more words. Visual features are extracted for each video segment, and textual features are extracted for each word. The visual features and the textual features are combined to generate a similarity matrix in which each element represents a similarity level between a respective video segment and a respective word. Segment-attended sentence features are generated for the textual query based on the textual features and the similarity matrix. The segment-attended sentence features are combined with the visual features of the video segments to determine a plurality of alignment scores, which is used to retrieve a subset of the video content associated with the textual query to be retrieved from the video segments.

Abstract: An aspect includes a computer-implemented method for surgical workflow visualization as deviations to a standard workflow. Surgical videos documenting multiple cases of a medical procedure are analyzed to identify variations when the medical procedure is performed multiple times by one or more service providers. According to some aspects, a visualization of one or more of the variations are output to a display device.

Abstract: The present application relates to a smoke detection method based on deep learning, a device and a storage medium. The method includes the following steps: performing image enhancement processing and color transfer processing on a smoke image to obtain a color transferred smoke image; superimposing the color transferred smoke image on an indoor image to obtain an initial image, and performing screening processing, detection frame updating processing, feature extraction and layer calibration processing on the initial image to obtain a smoke calibrated feature image; performing segmentation processing, detection frame prediction processing, detection frame classification processing and image classification processing on the smoke calibrated feature image in sequence to obtain a smoke image set; and screening the smoke image set to obtain a target smoke image, and outputting a smoke detection result. The present disclosure realizes the collection of smoke data sets and improves the detection accuracy of smoke.

Abstract: In one embodiment, an apparatus comprises a communication interface and a processor. The communication interface is to communicate with a plurality of cameras. The processor is to obtain metadata associated with an initial state of an object, wherein the object is captured by a first camera in a first video stream at a first point in time, and wherein the metadata is obtained based on the first video stream. The processor is further to predict, based on the metadata, a future state of the object at a second point in time, and identify a second camera for capturing the object at the second point in time. The processor is further to configure the second camera to capture the object in a second video stream at the second point in time, wherein the second camera is configured to capture the object based on the future state of the object.

Abstract: One aspect of the present disclosure relates to an analysis device including one or more memories and one or more processors. The one or more processors are configured to estimate an arrangement region of a group of products of a same type based on a sales floor image, and notify information on a display state of the group of products in the arrangement region estimated.

Abstract: A camera system may automatically identify high frequency motion regions in a field of view of the camera system. Analysis of motion events detected at different regions of a scene in the field of view, may be performed. Similar events frequently occurring at a certain region may be recorded/accumulated and compared to a threshold. Regions with high frequency activity may be determined and a user may be notified. If the user is not interested in events within the high activity regions, the regions may be ignored and notifications to the user may cease. The camera system may determine specific actions/events occurring in a region, such as a person approaching or leaving a home.

Abstract: A system for monitoring vehicle traffic includes one or more non-transitory computer-readable storage media storing computer-executable instructions that, when executed by one or more processors, cause a computing system to recognize, using images, at least one of a first face of a person in a facial detection zone, compare the first face to a database, identify the person as a suspicious person, and correlate the first face with a stored electronic device identifier. The computer-executable instructions further cause the computing system to access at least one personal identification database, and retrieve personal identification information of a suspicious individual by correlating information of the personal identification database with the first face or the stored electronic device identifier.

Abstract: Methods and systems are provided that include obtaining first camera images of a parking location before a vehicle is parked at the parking location; obtaining second camera images of the parking location after the vehicle is parked at the parking location; and detecting, via a processor, a fluid leak for the vehicle based on a comparison of the first camera images and the second camera images.

Abstract: A data processing system efficiently identifies a target object pointed to by a vehicle occupant. The data processing system provides a position detection unit that detects the position of the vehicle; an occupant state recognition unit that recognizes motion of an occupant pointing to outside of the vehicle; a target object database that indicates position of target objects that may be pointed to by an occupant; an appearance feature database that indicates appearance features of the target object; a speech recognition unit that recognizes words indicative of appearance features from the speech of the occupant; an object recognition unit that extracts target object candidates pointed at by the occupant by searching the target object database and the appearance feature database using a direction pointed at by an occupant, and a word included in the speech recognized by the speech recognition unit; and an output unit that outputs target object candidates.

Abstract: The present disclosure provides a processing method, a device and a storage medium for rear view image. The processing method includes following steps: obtaining an original rear view image; obtaining vehicle data to determine at least one basic curvature; obtaining driver information to determine at least one floating curvature; and processing the rear view image according to the at least one basic curvature and the at least one floating curvature. By performing these steps, the processing method can adjust display curvature of the rear view image according to specific driving scenario and user needs, thereby clearly and personalized displaying rear view image with suitable range and display scale on the display screen with limited area, to help the user clearly observe situations on the left, right and rear sides of the vehicle.

Abstract: The disclosure generally pertains to systems and methods for detecting obstacles in the vicinity of power cables. In an example method, a first end of a power cable may be detected to be connected to a vehicle. A first trajectory of the power cable may then be determined via a plurality of sensors at the vehicle. The first trajectory of the power cable may include a first location and a first orientation of the power cable. It may then be determined, via the plurality of sensors at the vehicle, whether an obstacle is within a predetermined distance of the first trajectory of the power cable. Responsive to the determination that the obstacle is within the predetermined distance of the first trajectory of the power cable, a notification associated with the obstacle may be provided, and a power supply to the power cable may be turned off.

Abstract: A system for collision detection includes an interface and a processor. The interface is configured to receive sensor data from a plurality of sensors. The processor is configured to preprocess the sensor data using a set of preprocessing steps to create preprocessed sensor data, wherein the set of preprocessing steps comprises filtering, normalization, and alignment; process the preprocessed sensor data using a compound model to create a collision score; and provide the collision score.

Abstract: Methods, systems, and apparatuses related to autonomous vehicle object detection are described. An autonomous vehicle can capture an image corresponding to an unknown object disposed within a sight line of the autonomous vehicle. Processing resources available to a plurality of memory devices associated with the autonomous vehicle can be reallocated in response to capturing the image and an operation involving the image corresponding to the unknown object to classify the unknown object can be performed using the reallocated processing resources.

Abstract: System, methods, and other embodiments described herein relate to the detection of lanes in a driving scene through segmenting road regions using an ontology enhanced to derive semantic context. In one embodiment, a method includes segmenting an image of a driving scene, independent of maps, by lane lines and road regions defined by an ontology and a pixel subset from the image has semantics of lane information from the ontology. The method also includes computing pixel depth from the image for the lane lines and the road regions using a model. The method also includes deriving 3D context using relations between the semantics and the pixel depth, the relations infer a driving lane for a vehicle from types of the lanes lines and the road regions adjacent to the driving lane. The method also includes executing a task to control the vehicle on the driving lane using the 3D context.

Abstract: Disclosed are an intersection recognition and restoration system using geometric modeling method and intersection recognition and restoration method using the same. The intersection recognition and restoration system using geometric modeling of the present invention for solving the above problems comprises an image capturing unit that sequentially acquires images from a front camera installed in a subject vehicle; an image analysis unit that detects a driving lane from a sequential image of the video file acquired by the image capturing unit, and recognizes and restores an intersection by a geometric modeling; a memory for storing driving lane and intersection data detected by the image analysis unit; and a display unit that expresses the driving lane and intersection data stored in the memory in the form of a top view.

Abstract: A method for recognizing a road intersection ahead in the driving direction of a vehicle. In the method, an overall probability of the presence of the road intersection ahead is calculated on the basis of data acquired during the trip of the vehicle and the presence of the road intersection ahead is deduced on the basis of the calculated overall probability.

Abstract: In various examples, systems and methods are disclosed herein for a vehicle command operation system that may use technology across multiple modalities to cause vehicular operations to be performed in response to determining a focal point based on a gaze of an occupant. The system may utilize sensors to receive first data indicative of an eye gaze of an occupant of the vehicle. The system may utilize sensors to receive second data indicative of other data from the occupant. The system may then calculate a gaze vector based on the data indicative of the eye gaze of the occupant. The system may determine a focal point based on the gaze vector. In response to determining the focal point, the system causes an operation to be performed in the vehicle based on the second data.

Abstract: The present disclosure relates to a method of matching a text with a design performed by an apparatus for matching a text with a design. According to an embodiment of the present disclosure, the method may comprise acquiring an image from information including images and texts; learning features of the acquired image; extracting texts from the information and performing learning about a pair of an extracted text and the acquired image; extracting a trend word extracted at least a predetermined reference number of times among the extracted texts; performing learning about a pair of the trend word and the acquired image; and identifying a design feature matched with the trend word among learned features of the image.

Abstract: A system for transmitting color and depth information, comprising: an image capturing device capturing a full color image frame of a scene; a depth capturing device capturing a grayscale depth image frame of the scene; a target identifying module for identifying a target in the scene; a bounding box generator configured for generating a bounding box encapsulating the target and for creating bound grayscale depth images and bound full color images; a depth colorizer converting the bound grayscale depth images into a bound full color depth image by linking and normalizing the depth dimension of the bounding box to a transferred color gamut by implementing a transfer function; an image merger that merges the bound full color images and bound full color depth images together into bound merged full color images; and an image encoder compressing the bound merged full color images into an image bitstring.

Abstract: A blood analyser and related methods, in particular a method of analysing a blood sample is disclosed. The blood analyser comprises a memory, an interface, and one or more processors. The blood analyser is configured to obtain image data of a prepared blood sample; select a first image associated with a first image plane of the prepared blood sample from the image data; characterize the first image, wherein the characterization of the first image comprises to determine a first set of cell regions belonging to the first image plane; select a first distal image associated with a first distal image plane on a distal side of the first image plane, and determine a first distal set of distal cell regions associated with the first set of cell regions; and determine a first cell parameter for each cell region of the first set of cell regions based on the first distal set of distal cell regions.

Abstract: In some examples, deterring a third party from copying a product or process may include scanning code to generate scanned code. Text in the scanned code may be read and/or at least one image in the scanned code may be recognized. The text and/or the at least one image may be analyzed to determine whether the code includes a specified trademark. Based on a determination that the code includes the specified trademark, an indication of authenticity of the code may be generated.

Abstract: In one embodiment, a system receives an image depicting a line of text. The system segments the image into two or more fragment images. For each of the two or more fragment images, the system determines a first hypothesis to segment the fragment image into a first plurality of grapheme images and a first fragmentation confidence score. The system determines a second hypothesis to segment the fragment image into a second plurality of grapheme images and a second fragmentation confidence score. The system determines that the first fragmentation confidence score is greater than the second fragmentation confidence score. The system translates the first plurality of grapheme images defined by the first hypothesis to symbols. The system assembles the symbols of each fragment image to derive the line of text.

Abstract: A document management system uses an objective loss function to improve the performance of optical character recognition (OCR) processes on images of documents. A user captures an image of a document, which is imported into the document management system. The document management system performs OCR on a high resolution version of the image of the document, obtaining a first set of text representative of the text of the document. The document management system applies a machine-learned model on a low-resolution version of the image of the document, producing a denoised image that is of a higher resolution than that input into the machine-learned model. The document management system performs OCR on the denoised image, obtaining a second set of text representative of the text of the document. The document management system compares the first and second sets of text in the form of an objective loss function.

Abstract: Computer-implemented methods of automatically identifying, ranking, and processing information obtained from a document, and computerized systems and computer program products related thereto. The method involves identifying text clusters and identifying a visual layout structure of at least one part of the document, and ranking the text clusters according to visual properties of the text cluster. The method further involves identifying a semantic context of the identified text clusters and ranking the text clusters according to a similarity of the identified semantic context in relation to a given semantic context, to obtain a semantic context ranking, creating a total ranking of the text clusters based on a combination of a pair of rankings, and selecting text cluster(s) according to its position in the total ranking and providing the selected text cluster(s) to at least one downstream application.

Abstract: A device configured to receive a request for processing a digital document that identifies a target data structure. The device is further configured to obtain the digital document and to process the digital document to identify objects within the digital document. The device is further configured to input the digital document into a machine learning model that is configured to output contextual metadata for the digital document based on the digital document. The contextual metadata describes attributes of the identified objects within the digital document. The device is further configured to populate a first subset of entries in the target data structure with identifiers for the objects identified within the digital document and to populate a second subset of entries in the target data structure with the contextual metadata. The device is further configured to send the populated target data structure to the network device.

Abstract: A computing device may access visually rich documents comprising an image and metadata. A graph, based on the image or metadata, can be generated for a visually rich document. The graph's nodes can correspond to words from the visually rich document. Features for nodes can be determined by the device. The device may generate model labeled graphs by assigning a pseudo-label to nodes using a pretrained model. The device may generate a plurality of graph labeled graphs by assigning a pseudo-label to nodes by matching a first node from a first graph to at least a second node from a second graph. The device may generate a plurality of updated graphs by cross referencing labels from the model labeled graphs and the graph labeled graphs. Until a change in labels is below a threshold, a model can be trained to perform key-value extraction using the updated graphs.

Abstract: Described herein is a system for automatically capturing images. A device may execute an application on the device. In response to receiving a request for capturing an image of a document, the application may control the operation of the camera to contemporaneously and continuously capture image frames of an area within the field of view of a camera of the device. The field of view may include an object and in turn may be included in the captured image frames. The application may determine whether the aspect ratio and real-world size of the object in the captured image frames corresponds with a specified threshold size and aspect ratio which corresponds with a given document type. In response to determining the object corresponds to the given document type, the camera may automatically extract an image of the object.

Abstract: Novel tools and techniques are provided for implementing automatic document pattern recognition and analysis. In various embodiments, a computing system might receive one or more first documents containing first remittance information. The computing system might analyze the one or more first documents to extract first remittance information from the first document by obtaining one or more historical documents and correlating the one or more first documents to the one or more historical documents. Next, based on the correlation of the one or more first documents to the one or more historical documents, the computing system might determine at least one historical document that correlates to the one or more first documents. Additionally, the computing system might extract the first remittance information from the one or more first documents based on the at least one historical document.

Abstract: The present invention realizes an information processing device capable of accurately differentiating an object from within an imaged image regardless of imaging conditions. The present invention also realizes an information processing method and a control program. A first circumscribed polygon region calculation unit (32) calculates a first circumscribed polygon region as a convex hull circumscribed region that contains a feature region. A principal axis of inertia calculation unit (33) calculates a principal axis of inertia for the feature region. A region division unit (34) uses the principal axis of inertia to divide the feature region. On the basis of the area ratio of the two second circumscribed polygon regions and the first circumscribed polygon region as convex hull circumscribed regions that respectively contain the divided feature regions, a region selection unit (36) selects either the first circumscribed polygon region or the two second circumscribed polygon regions.

Abstract: An apparatus for performing image analysis to identify human actions represented in an image, comprising: a joint-determination module configured to analyse an image depicting one or more people using a first computational neural network to determine a set of joint candidates for the one or more people depicted in the image; a pose estimation module configured to derive pose estimates from the set of joint candidates that estimate a body configuration for the one or more people depicted in the image; and an action-identification module configured to analyse a region of interest within the image identified from the derived pose estimates using a second computational neural network to identify an action performed by a person depicted in the image.

Abstract: A palm recognition device is provided, and belongs to the field of the electronic device technologies. The palm recognition device includes a housing, a PCB, and a camera. The PCB and the camera are both located inside the housing. The housing has an image capturing window, and a projection of the camera on the image capturing window deviates downward relative to a central point of the image capturing window.

Abstract: An apparatus may include an ultrasonic sensor stack, a foldable display stack and a transmission enhancement layer. The foldable display stack may include a display stiffener and display stack layers. The display stack layers may form one or more display stack resonators configured to enhance ultrasonic waves transmitted by the ultrasonic sensor stack in a first ultrasonic frequency range. In some implementations, a transmission enhancement resonator may include the display stiffener and the transmission enhancement layer. In some examples, the transmission enhancement resonator may include at least a portion of the ultrasonic sensor stack. The transmission enhancement resonator may be configured to enhance the ultrasonic waves transmitted by the ultrasonic sensor stack in the first ultrasonic frequency range.

Abstract: A display panel and an electronic device are provided. The display panel includes a first panel, a second panel, and a binding layer. The first panel includes light-emitting units, and also includes light-transmitting regions. The second panel includes photoelectric sensing elements and a second driving circuit layer. The second driving circuit layer includes second driving transistors. The second driving transistors are electrically connected to the photoelectric sensing elements, and the photoelectric sensing elements are opposite to the light-transmitting regions. The binding layer is disposed between the first panel and the second panel, and configure to bind the first panel and the second panel.

Abstract: Provided are a display panel and a display device. The display panel has an optical identification region and includes a first layer and a second layer. The first layer is located at a side of the second layer close to a light-exiting side of the display panel. The first layer has at least one first light-shielding region and includes at least one first light-transmitting hole. The second layer includes at least one second light-transmitting hole corresponding to the at least one first light-transmitting hole. One of the at least one first light-shielding region and one of the at least one second light-transmitting hole at least partially overlap with each other along a direction perpendicular to a plane of the display panel.

Abstract: A display device includes an optical fingerprint sensor; a display panel; a patterned film disposed below the display panel and having an opening portion that exposes the optical fingerprint sensor; a heat dissipating layer disposed below the patterned film, and that dissipates heat; and a cushion layer disposed between the patterned film and the heat dissipating layer, and that absorbs an impact. The optical fingerprint sensor is disposed below the cushion layer, and the heat dissipating layer has an opening portion that has a size larger than a size of the opening portion of the patterned film.

Abstract: An information processing device includes: an acquisition unit that acquires a rolled print and a slap print on the same finger; an extraction unit that extracts first feature points from the rolled print and second feature points from the slap print, respectively; and an estimation unit that, based on a ratio occupied by the first feature points having no correspondence with the second feature points out of the first feature points included in a common region that is common to the rolled print and the slap print, estimates a first work amount of correction work performed by a user on a non-common region excluding the common region from the rolled print.

Abstract: This application relates to a laptop computer. The laptop computer includes a base portion pivotally coupled to a lid portion is described. The laptop computer includes a display assembly carried by the lid portion, where the display assembly includes a light-transmissive cover, a display layer overlaid by the light-transmissive cover, a display stack electrically coupled to and overlaid by the display layer, and a light pattern recognition module adjacent to the display stack and overlaid by the display layer. The light pattern recognition module includes (i) a light pattern projector that projects a light pattern directly through the display layer.

Abstract: This disclosure provides methods, devices, and systems for object detection in images. The present implementations more specifically relate to cascaded object detection techniques. In some aspects, an object detection system may obtain a first image and perform a first object detection operation on the first image. The object detection system may select a region of interest (ROI) in the first image based at least in part on the object detection operation, the ROI being selected before the object detection operation is completed. The object detection system may obtain a second image coinciding with the ROI of the first image, the second image having a higher resolution than the first image. The object detection system may selectively perform an inferencing operation on the second image based on a result of the object detection operation.

Abstract: A recording medium stores a program for causing a computer to execute processing including: acquiring images; classifying the images, based on a combination of whether an action unit related to a motion of a portion occurs and whether occlusion is included in an image in which the action unit occurs; calculating a feature amount of the image by inputting each classified image into a model; and training the model so as to decrease a first distance between feature amounts of an image in which the action unit occurs and an image with an occlusion with respect to the image in which the action unit occurs and to increase a second distance between feature amounts of the image with the occlusion with respect to the image in which the action unit occurs and an image with an occlusion with respect to an image in which the action unit does not occur.

Abstract: Disclosed are system, method, and computer program product embodiments for determining a digital marker indicative of a neurological condition. An embodiment operates by detecting a region of interest in an image frame of a face of a user. The region of interest includes an eye of the user and the image frame is obtained from a video stream. The embodiment then generates a point spread function (PSF) based on acquired data and a portion of the image frame specified by the region of interest, performs deconvolution on the portion of the image frame specified by the region of interest using the PSF, determines a set of oculometric parameters of the eye of the user using the deconvolved image, and determines one or more digital markers of the user based on the set of oculometric parameters. The one or more digital markers are indicative of a neurological condition of the user.

Abstract: The present disclosure detects a pupil of an eye using circle formation based scoring method. The conventional approaches fail to provide an accurate and reliable biometric authentication due to the usage of simple thresholding based statistical methods and iris dependent segmentation methods. The present disclosure utilizes a circle plotting approach and selects the optimum circle using several parameters. The present disclosure can generate a pupil boundary that fits the pupil region inside an iris perfectly. Initially, the system receives an input image of an eye. After removing reflections, a core point of the reflection free image is identified. Further, a plurality of points are obtained based on a sudden gradient change. and a plurality of circles are plotted. Further, an optimum circle is identified using a score based optimum circle selection method. Finally, the pupil associated with the input image is identified based on the optimum circle.

Abstract: An iris recognition apparatus is provided with: an irradiating device that is capable of irradiating light to a target person to be recognized; an acquiring device that acquires a first image which imaged a first imaging range including at least a face of the target person; and a determining device that determines an irradiation mode of light irradiated from the irradiating device to the target person when imaging a second image including an area, which includes an eye of the target person, and corresponding to a second imaging range, which is narrower than the first imaging range, based on the first image.

Abstract: An iris recognition apparatus is provided with: an irradiating device that is capable of irradiating light to a target person to be recognized; an acquiring device that acquires a first image which imaged a first imaging range including at least a face of the target person; and a determining device that determines an irradiation mode of light irradiated from the irradiating device to the target person when imaging a second image including an area, which includes an eye of the target person, and corresponding to a second imaging range, which is narrower than the first imaging range, based on the first image.

Abstract: An information processing device including a processor configured to perform a first processing related to a position of at least one of a person and an object around the person using a positioning application and perform a second processing to determine an amount of food eaten by the person using a meal application, wherein the processor is configured to change an operation mode of the positioning application from a first mode to a second mode with a higher processing load of the positioning application than the first mode if the meal application determines at least one of the amount of the food or an amount of nutrients is insufficient.

Abstract: An emotion recognition method includes the following steps: acquiring video data of a human-object interaction behavior process; performing data labeling on the positions of a person and an object and the interaction behaviors and emotions expressed by the person; constructing a feature extraction model based on deep learning, extracting features of interaction between the person and the object in a time-space dimension, and detecting the position and category of the human-object interaction behavior; mapping the detected interaction behavior category into a vector form through a word vector model; and finally, constructing a fusion model based on deep learning, fusing the interaction behavior vector and the time-space interaction behavior features, and identifying the emotion expressed by the interaction person.

Abstract: A video processing method for performing partial highlighting with the aid of hand gesture detection and an associated SoC are provided. The SoC includes a person recognition circuit, a hand gesture detection circuit, a sound detection circuit and a processing circuit. The person recognition circuit obtains image data from an image capturing device, and performs person recognition on the image data to generate a recognition result. The hand gesture detection circuit performs hand gesture detection on hand gesture image data to generate a hand gesture detection result. The sound detection circuit receives multiple sound signals from multiple microphones, and determines a voice characteristic value of a main sound. The processing circuit determines a specific region in the image data according to the recognition result, the hand gesture detection result, and the voice characteristic value, and processes the image data to highlight the specific region.

Abstract: An information processing apparatus (2) includes an acquisition unit (211) configured to acquire a person image (IMG) including an eye area (EA) of a target person, and a determination unit (212) configured to determine that the target person impersonates another person in the case where an aperture (AP) enclosing at least a part of an eye of the target person is present in the person image.

Abstract: Provided are a method, device, and system for generating a liveness detection dataset, training a liveness detection model, and performing liveness detection. The method includes: obtaining a plurality of real images of faces; providing the plurality of real images to a neural network; generating a plurality of synthetic images corresponding to the plurality of real images based on an output of the neural network; and training a liveness detection model based on the plurality of real images and the plurality of synthetic images, wherein the liveness detection model is used to perform liveness detection by determining whether an input image of a face includes a live image of the face.

Abstract: A system and method including receiving an indication of an occurrence of an event associated with the operation of an autonomous vehicle; determining at least one action to be performed, the at least one action including a data request for a specified subset of stored data associated with the operation of the autonomous vehicle from a memory; generating an output including the specified subset of data; and transmitting the specified subset of data to a remote monitoring system.

Abstract: Aspects of the disclosure relate to a dynamic processing system for roadside service control and output generation. A computing platform may receive, from a client device, video content corresponding to a disabled vehicle, which may include geotagging information corresponding to a location of the disabled vehicle. Based on the video content and the geotagging information, the computing platform may determine a provider output indicating a potential service provider for assisting with the disabled vehicle. The computing platform may send, to the client device, an indication of the provider output. In response to receiving an indication that the potential service provider is acceptable, the computing platform may send a request to dispatch a driver of the potential service provider to the location of the disabled vehicle.