Abstract: A method and an apparatus for pattern recognition is provided in the present invention, applied to the field of artificial intelligence. The method includes: acquiring a two-dimensional image of a target object and a two-dimensional feature of the target object according to the two-dimensional image of the target object; and acquiring a three-dimensional image of the target object and a three-dimensional feature of the target object according to the three-dimensional image of the target object; identifying the target object according to the two-dimensional feature and the three-dimensional feature of the target object. The method can reduce restrictions on acquiring the image of the target object, for example, reduce the restrictions on the image of the target object in terms of postures, lighting, expressions, make-up and occlusion, thereby improving an accuracy of recognizing the target object and improving a recognition rate and reducing recognition time at the same time.

Abstract: A device and method for achieving a high probability of match for identifying individuals in a particular group, is provided. The device and method consists of two stages. The first stage compares a biometric template against personnel in an authorized group with a high probability of match standard. In the second stage, in response to no match being made in the authorization group, a search would be conducted against a second group, such as a watch list, with the same or lower probability of match rate.

Abstract: A computer implemented method for gender classification by applying feature learning and feature engineering to face images. The method includes conducting feature learning on a face image comprising feeding the face image into a first convolution neural network to obtain a first decision, conducting feature engineering on a face image, comprising the steps of automatically detecting facial landmarks in the face image, transforming the facial features into a two-dimensional matrix, and feeding the two-dimensional matrix into a second convolution neural network to obtain a second decision, computing a hybrid decision based on the first decision and the second decision, and classifying gender of the face image in accordance with the hybrid decision.

Abstract: A method for recognizing, identifying, and authenticating persons to be identified from their faces in a preset scenario, includes uploading a first image of users to a server, and receiving the first image on which first facial rectangles and user information have been imposed, and capturing a second image and detecting second facial rectangles therefrom. The method determines whether a number of faces of the second image is equal to a number of faces of the first image, and uploading the second image to the server for identifying third facial rectangles of the second image. A plurality of distances is obtained by calculating at least one distance between third and second facial rectangles of each user shown in the second image and determining whether the users shown in the second image are the same as users shown in the first image based on the plurality of distances.

Abstract: A management system for supervising an operator within an operation area with a processing apparatus installed therein has an imaging camera for capturing an image of the operator and a control unit. The control unit stores an area map that includes information about the processing apparatus and the positions of the processing apparatus, acquires identifying information regarding the operator on the basis of the image captured by the imaging camera, and extracts traffic lines of the operator. A management unit records the traffic lines of the operator and the identifying information of the operator, in relation to each other on the area map.

Abstract: Methods, devices and systems for training a pattern recognition system are described. In one example, a method for training a sign language translation system includes generating a three-dimensional (3D) scene that includes a 3D model simulating a gesture that represents a letter, a word, or a phrase in a sign language. The method includes obtaining a value indicative of a total number of training images to be generated, using the value indicative of the total number of training images to determine a plurality of variations of the 3D scene for generating of the training images, applying each of plurality of variations to the 3D scene to produce a plurality of modified 3D scenes, and capturing an image of each of the plurality of modified 3D scenes to form the training images for a neural network of the sign language translation system.

Abstract: An image processing apparatus includes a first detection unit that executes first detection processing for detecting an object presumed to be a human from an image by using a trained model relating to humans, a second detection unit that executes second detection processing for detecting an object presumed to be a human from the image at least by using pattern matching, a determination unit that determines whether the object detected through the first detection processing corresponds to the object detected through the second detection processing, a correction unit that corrects a detection result relating to the object that is detected through the first detection processing and is determined to correspond to the object detected through the second detection processing, and a calculation unit that calculates the number of humans in the image based at least the detection result corrected by the correction unit.

Abstract: Described herein is a human pose prediction system and method. An image comprising at least a portion of a human body is received. A trained neural network is used to predict one or more human features (e.g., joints/aspects of a human body) within the received image, and, to predict one or more human poses in accordance with the predicted one or more human features. The trained neural network can be an end-to-end trained, single stage deep neural network. An action is performed based on the predicted one or more human poses. For example, the human pose(s) can be displayed as an overlay with received image.

Abstract: In one aspect, a computerized method useful for handwriting recognition (HWR) on physical mail envelopes addressed to a user includes the step of scanning a physical mail item to obtain a digital image of the address-side of the physical mail item. The method includes the step of identifying that at least one of a return address region or a recipient address region of the address-side of the physical mail item. The method includes the step of determining that the at least one of the return address region or a recipient address region comprises a handwritten text. The method includes the step of providing a data store of known senders to the recipient address. The data store of known senders comprises a data store of return address information in a known sender handwriting samples and a data store of receiver address information in the known sender handwriting samples. The method includes the step of providing a data store of a receiver's identity and address.

Abstract: An identity recognition device, and an substrate and an identity recognition circuit are disclosed. In the substrate, a plurality of second recognition output lines are added to be arranged in pairs with first recognition output lines. A control signal loaded to the control signal line will be coupled to the first and second recognition output lines by the parasitic capacitance between the control signal line and the recognition output lines to cause similar signal interference to the first and second recognition output lines. During the identity recognition time period, the signal interference in the electric signal outputted by the first recognition output line can be removed by utilizing the similar noise signal received by the second recognition output line, in such a way to improve the signal-to-noise ratio and detection precision of the identity recognition signal.

Abstract: The present invention provides a method for separating out a defect image from a thermogram sequence based on weighted naive Bayesian classifier and dynamic multi-objective optimization, we find that different kinds of TTRs have big differences in some physical quantities. The present invention extracts these features (physical quantities) and classifies the selected TTRs into K categories based on their feature vectors through a weighted naive Bayesian classifier, which deeply digs the physical meanings contained in each TTR, makes the classification of TTRs more rational, and improves the accuracy of defect image's separation. Meanwhile, the multi-objective function does not only fully consider the similarities between the RTTR and other TTRs in the same category, but also considers the dissimilarities between the RTTR and the TTRs in other categories, thus the RTTR selected is more representative, which guarantees the accuracy of describing the defect outline.

Abstract: A device, system and method for generating an alert and an automatic search for a candidate subject is provided. The device monitors one or more data sources to identify a candidate subject. The device determines, from rules for rating candidate subjects stored in a memory, a rating of the candidate subject. When the rating meets a first threshold: the device transmits, to one or more communication devices, an alert identifying the candidate subject. When the rating meets a second threshold, lower than the first threshold: the devices provides, a classifier defining the candidate subject to an object recognition engine to cause the object recognition engine to automatically search for the candidate subject, using the classifier, in multimedia data received at one or more edge devices. The multimedia data may include images.

Abstract: An information processing method including the following executed using a computer: obtaining a neural network model that solves a regression problem; obtaining input data and label data corresponding to the input data; compressing a network of the neural network model to obtain a compressed model; transforming the regression problem to be solved by the neural network model into a classification problem, based on information indicating performance of the compressed model, the information being calculated using the label data and output data which is obtained by inputting the input data to the compressed model; and changing a network configuration of the neural network model and transforming the label data, in accordance with the transformation from the regression problem to the classification problem.

Abstract: A computing system determines if an event has occurred. A first window is defined that includes a subset of a plurality of observation vectors modeled as an output of an autoregressive causal system. A magnitude adjustment vector is computed from a mean computed for a matrix of magnitude values that includes a column for each window of a plurality of windows. The first window is stored in a next column of the matrix of magnitude values. Each cell of the matrix of magnitude values includes an estimated power spectrum value for a respective window and a respective frequency. A second matrix of magnitude values is updated using the magnitude adjustment vector. Each cell of the second matrix of magnitude values includes an adjusted power spectrum value for the respective window and the respective frequency. A peak is detected from the next column of the second matrix of magnitude values.

Abstract: A method of detecting clouds in an acquired aerial image includes determining a region of a reference aerial image corresponding to a region of an acquired aerial image. For each of a plurality of locations over the region of the acquired aerial image and corresponding to a plurality of locations over the region of the reference aerial image, the mutual information of one or more variables associated with the location in the acquired aerial image and one or more variables associated with the corresponding location in the reference aerial image is calculated. Using the mutual information calculated for each of the plurality of locations over the region of the acquired aerial image, it is determined when the acquired aerial image displays a cloud at the location in the region of the acquired aerial image.

Abstract: The present subject matter provides a technical solution for various technical problems associated with precision agriculture imagery. A technical solution for stitching together successive passes of captured images may include using multiple image capture devices, where the image capture devices have different fields of view. A first camera may be used to provide a first field of view (FOV), and may be configured to provide a selected ground sample distance (GSD) for precision agriculture imagery analysis. A second camera may be used to provide a wider FOV, and may be used to provide sufficient stitching process overlap between images captured in successive passes.

Abstract: In one embodiment, a technique includes detecting an identifier associated with a product. The technique further includes accessing a content object associated with the identifier on a third-party system. In the technique, the content object includes a set of instructions. The technique also includes detecting a trigger. The technique further includes selecting a step of the content object and outputting the selected step.

Abstract: A head-mounted display device interfaces with a medical device configured to perform an invasive procedure on a patient. The display device includes a frame for mounting on a person's head, a display, a wireless transceiver configured to communicate with a network, and a processing circuit. The processing circuit is coupled to the frame, the display and the wireless transceiver and receives input data relating to the medical device. The processing circuit retrieves from a memory an instruction relating to the medical device based on the input data and displays the instruction relating to the medical device on the display.

Abstract: Systems, methods, and non-transitory computer-readable media can receive a geometric map and a semantic map associated with a geographic area, the semantic map comprising semantic data associated with vehicle navigation. A first semantic position estimate associated with a first piece of semantic data contained in the semantic map is generated based on semantic data location information associated with the first piece of semantic data. A final position for the first semantic position estimate is received. One or more three-dimensional semantic labels are applied to the geometric map based on the final position of the first semantic position estimate. A warped semantic map is generated. Generating the warped semantic map comprises warping the semantic map based on the one or more three-dimensional semantic labels.

Abstract: This disclosure describes an artificial reality system that presents artificial reality content in the context of a physical environment that includes a mirror or other reflective surface. In one example, this disclosure describes a method that includes capturing capture data representative of a physical environment, wherein the physical environment includes a reflective surface and a plurality of objects, determining a pose of the HMD, determining a map of the physical environment, wherein the map includes position information about the reflective surface and position information about each of the plurality of physical objects in the physical environment, identifying a visible object from among the plurality of physical objects, and generating artificial reality content associated with the visible object.

Abstract: A cognitive load reducing platform comprises one or more sensors that collect information about an environment as sensor data. A processor is in communication with the one or more sensors, wherein the processor executes a cognitive enhancement engines that processes the sensor data into enhanced characterization data having a reduced amount of data compared to the sensor data. An output device electronically communicate the enhanced characterization data to a user such that the enhanced characterization data reduces the cognitive load of the user, wherein the sensor and the output device comprise an assisted perception module worn by the user during an incident. A command interface device remote from the user to enable a person of authority to manage the incident and the user by receiving and displaying the enhanced characterization data, and by transmitting data and commands back to the assisted perception module.

Abstract: Aspects of the present disclosure involve a system comprising a computer-readable storage medium storing a program and a method for detecting a pose of a user. The program and method include operations comprising receiving a monocular image that includes a depiction of a body of a user; detecting a plurality of skeletal joints of the body based on the monocular image; accessing a video feed comprising a plurality of monocular images received prior to the monocular image; filtering, using the video feed, the plurality of skeletal joints of the body detected based on the monocular image; and determining a pose represented by the body depicted in the monocular image based on the filtered plurality of skeletal joints of the body.

Abstract: A target identification method includes: using information of a to-be-detected target acquired within a predetermined time period as judgment information; acquiring an identification result of the to-be-detected target at a current time and outputting the identification result; judging whether the attribute type corresponding to the identification result is an attribute type having the highest priority; and if the attribute type corresponding to the identification result is not the attribute type having the highest priority, using information of the to-be-detected target acquired within a next predetermined time period as the judgment information, and returning to the step of acquiring an identification result of the to-be-detected target at a current time and outputting the identification result.

Abstract: The present disclosure describes techniques for identifying images of users in visual media data. One example method includes receiving visual media data from a client device, wherein the visual media data was captured in real time by the client device; detecting a presence of an image of a human in the visual media data; determining a user identification associated with the human based on the detected presence of the image of the human in the visual media data; retrieving data corresponding to the user based on the user identification; generating user graphics data based on the retrieved data corresponding to the user; and transmitting the user graphics data to the client device, wherein the client device is configured to display the user graphics data with the visual media data.

Abstract: A method and system for producing video-segments of a live-action event involving monitoring a live-action event for detection of event-segments, detecting one or more event-triggers with detectors, determining if an event-segment occurred based on the detected event-triggers, and editing one or more video feeds into a video-segment to encompass the event-segment.

Abstract: A monitoring system includes a first imaging device, a controller, storage, and a notifying section. The first imaging device captures an image of a first imaging area to generate first captured image data indicating a first captured image. The controller detects the events based on the first captured image data. The storage stores data indicating a first flag and a second flag. The events include a pre-event and a post-event detected after the pre-event has ended. The post-event indicates a detection target removal act for a detection target placed in the first imaging area. The controller changes the first flag from the off state to an on state upon detecting the pre-event, holds the first flag in the on state even after the pre-event has ended, changes the second flag from the off state to the on state upon detecting the post-event, and notifies of consecutive event occurrence information.

Abstract: A system may receive video data generated by a camera device positioned in a physical environment. The system may receive electromagnetic (EM) response signals generated by an array of 5G antenna devices. The system may refine the EM response signals. The system may determine location information for a metallic object. The system may generate a movement trajectory of the metallic object. The system may generate a feature vector comprising an aggregation of the refined EM response signals and the movement trajectory. The system may submit the feature vector to a machine learning framework. The system may classify the metallic object as a weapon. The system may generate a virtual render comprising a graphical indication of the metallic object. The system may combine the virtual render with the video input data and display the combined virtual render and video data.

Abstract: A computing system automatically detects, in a sequence of video frames, a video frame region that depicts a scoreboard. The video frames of the sequence depict image elements including (i) scoreboard image elements that are unchanging across the video frames of the sequence and (ii) other image elements that change across the video frames of the sequence. Given this, the computing system (a) receives the sequence, (b) engages in an edge-detection process to detect, in the video frames of the sequence, a set of edges of the depicted image elements, (c) identifies a subset of the detected set of edges based on each edge of the subset being unchanging across the video frames of the sequence, and (d) detects, based on the edges of the identified subset, the video frame region that depicts the scoreboard.

Abstract: The present disclosure discloses a method and apparatus for determining video material of news. The method for determining video material of news comprises: recognizing a person name in a news text; searching a video based on the person name, to obtain a to-be-selected video; extracting a key frame in the to-be-selected video; recognizing a person in the key frame to obtain identity information of the person; and determining the to-be-selected video as video material of news, in response to the identity information of the person conforming to the person name. The present disclosure improves the consistency between the video material of the news and the news text.

Abstract: A system, method and program storage device are provided for automatically associating evidence recorded by a plurality of cameras with a discrete occurrence, including: receiving occurrence data pertaining to the discrete occurrence and storing at least a portion of the occurrence data in an occurrence record; receiving first evidence data comprising at least a video data portion and a metadata portion of the evidence recorded by a first camera of the plurality of cameras and storing it in an evidence record; receiving second evidence data comprising at least a video data portion and a metadata portion of the evidence recorded by a second camera of the plurality of cameras and storing it in the evidence record; automatically associating information stored in the evidence record with information stored in the occurrence record based on a correspondence of at least two criteria including a first criterion of time; identifying, based on the automatic association, a first image data portion of the evidence recor

Abstract: A smart surveillance system includes a communicator configured to communicate with a closed circuit television (CCTV) camera and a cloud server, a background image extractor configured to analyze CCTV image data received from the CCTV camera and to extract a background image, an object image analyzer configured to distinguish an object image from the background image through big data analysis of a CCTV image, a region of interest (ROI) extractor configured to extract an ROI corresponding to the object image, and a controller configured to provide the background image and the ROI to the cloud server, to receive the background image and the ROI from the cloud server, to combine the background image and the ROI, and to generate a complete CCTV image, if necessary.

Abstract: An autonomous vehicle is described herein. The autonomous vehicle includes several different types of sensor systems, such as image, lidar, and radar. The autonomous vehicle additionally includes a computing system that executes several different object classifier modules, wherein the object classifier modules are configured to identify types of objects that are in proximity to the autonomous vehicle based upon outputs of the sensor systems. The computing system additionally executes a Bayesian object classifier system that is configured to receive outputs of the object classifier modules and assign labels to objects captured in sensor signals based upon the outputs of the object classifier modules.

Abstract: The present disclosure provides perception system for a vehicle that includes a plurality of imaging devices for producing images of an environment of the vehicle; a perception filter for receiving the images produced by the imaging devices, wherein the perception filter crops and filters the received images based on an intent of the vehicle and a current state of the vehicle; and a perception module for receiving at least one of the cropped and filtered images from the perception filter and perceiving the environment of the vehicle based on the received at least one of the cropped and filtered images.

Abstract: The present disclosure provides a perception system for a vehicle. The perception system includes a number of imaging devices associated with the vehicle and a perception filter for determining a region of interest (“ROI”) for the vehicle based on an intent of the vehicle and a current state of the vehicle. The ROI is used to select images of an environment of the vehicle produced by the imaging devices and the perception filter receives and processes the images produced by the imaging device. The perception system further includes a perception module for receiving the processed images from the perception filter and perceiving the environment of the vehicle based at least in part on the received images.

Abstract: Calibration and distance prediction for driving assistance is provided. A camera of a vehicle is calibrated to obtain a distance data set. The distance data set includes a distance of each row of pixels of a first image captured by the camera. The distance data set may be further utilized in real-time to predict a distance of an object from the vehicle. Based on the predicted distance, a warning message for an impending collision may be generated and communicated to a driver of the vehicle, thereby facilitating driving assistance to the driver in the real-time.

Abstract: A method for detecting an obstacle applicable in an electronic device includes detecting whether at least one object is within a line of sight of an image capturing device. The image capturing device is controlled to capture a first image of the object and the image capturing device is caused to move until a capturing angle for capturing another image of the object is changed. The image capturing device is controlled to capture a second image of the object and a determination is made as to whether the object in the first image is the same as the object in the second image. For such recognized objects, the object is determined to be a non-planar obstacle when the object in the first image is not the same as the object in the second image.

Abstract: A communication device is provided for a motor vehicle that works with traffic-relevant information provided through light signals. The communication device comprises a primary imaging optical system to create the input image from the received scenic image, which may comprise a light car2car signal and/or a light car2X signal including traffic-relevant information. It further comprises a DMD array of mirrors for receiving the input image and transmitting it to a primary optical sensor adapted to convert the input optical image to an electric signal for processing, to thereby find out whether and what places of the input image sources of the signals are found in.

Abstract: The disclosure discloses a method and apparatus for identifying a traffic light. An embodiment of the method comprises: zooming a to-be-processed image acquired by an image acquisition device by at least one preset ratio to obtain at least one zoomed image; inputting the at least one zoomed image into a pre-trained convolutional neural network to obtain location information and category information of a traffic light corresponding to each zoomed image of the at least one zoomed image, wherein the convolutional neural network is used for retrieving location information and category information of a traffic light displayed in an image; and analyzing the obtained location information and category information to generate at least one candidate traffic light identification result, and fusing the generated candidate traffic light identification result to generate a traffic light identification result corresponding to the to-be-processed image. The embodiment improves the accuracy in identifying a traffic light.

Abstract: An occupant detection device of an embodiment includes, for example: an acquirer that acquires captured image data output from an image capturer capturing a seat and distance data indicating a distance to an occupant present in a direction of the seat; a seat specifier that specifies a position of a surface of the seat; a trunk surface specifier that specifies a position of a surface representing a trunk of the occupant based on the captured image data and the distance data; and a calculator that calculates a volume of the trunk of the occupant present between the seat and the image capturer based on the position of the surface representing the trunk and the position of the surface of the seat.

Abstract: A biometric device includes a substrate, an image sensor, an optical layer and at least one infrared light emitting diode (IR LED). The image sensor is disposed on the substrate. The optical layer is disposed on the image sensor and includes a diffraction pattern. The IR LED is disposed on the diffraction pattern of the optical layer. The optical layer is located between the IR LED and the image sensor.

Abstract: Systems and methods for repairing cross-linked biometric records receive a set of biometric records. Each biometric record contains at least one biometric sample in a non-textual modality. One or more of the biometric records in the set of biometric records is potentially a cross-linked biometric record having at least two biometric samples that are associated with different individuals. Crosslink resolution is performed on the set of biometric records by searching for a match between a biometric sample in a given non-textual modality of a given biometric record with each biometric sample of the same given non-textual modality in each of the other biometric records in the set of biometric records. During the crosslink resolution, biometric sample may be removed from one biometric record and merged with another biometric record.

Abstract: A system provides intelligent gallery management for biometrics. A first gallery is obtained that includes biometric and/or other information on a population of people. An application is identified. A subset of the population of people is identified based on the application. A second gallery is derived from the first gallery by pulling the information for the subset of the population of people without pulling the information for the population of people not in the subset. Biometric identification (such as facial recognition) for the application may then be performed using the second gallery rather than the first gallery. In this way, the system is improved as less time is required for biometric identification, fewer device resources are used, and so on.

Abstract: Embodiments relate to a two-stage end-to-end text recognition system. The text recognition system includes a text detection stage and a text recognition stage. Images inputted to the text recognition system are provided to both the text detection stage and to the text recognition stage. The text detection stage detects text regions in the images and provides the detected regions to the text recognition stage. The text recognition stage is trained to perform geometric rectification on the text regions using the images. There is end-to-end alignment between the text detection stage and the text recognition stage. Additionally, the text detection stage and text recognition stage are each trained independent of the other.

Abstract: A method for acquiring images of the surface of the earth, installing an aerial platform in a quasi-stationary position, equipped with an image acquisition system with a large field of view and a second, high-resolution, image acquisition system is disclosed. The method includes implementing successive observation cycles, each one including the acquisition of an image of a zone of interest by the first system, the partitioning of the image thus acquired into mesh units which each correspond to a sector of the zone of interest, the analysis of the image in order to detect the potential presence of unwanted marks, and the acquisition of an image by the second system for the mesh units for which no unwanted marks have been detected. Observation cycles are thereby implemented until images of the entire zone of interest have been acquired by the second system.

Abstract: An image acquisition system includes a source of a radiation, an image sensor including an array of photodetectors capable of detecting the radiation and including a surface, and an angular filter, covering the sensor, and capable of blocking the rays of the radiation having an incidence relative to a direction orthogonal to the surface greater than a threshold and of giving way to at least certain rays of the radiation having an incidence relative to a direction orthogonal to the surface smaller than the threshold.

Abstract: Provided are a camera and an image processing method of the camera. According to an aspect of the present disclosure, an image processing method of a camera includes receiving region information; receiving image data extracted from a corresponding region of an original resolution image corresponding to a region set according to the region information regarding an input image; and analyzing image data of the original resolution image and detecting an object corresponding to a set condition.

Abstract: With the present invention, it is possible to swiftly collect light source information for use when correcting color information of an object region in a captured image. A position in real space of an object included in a captured image is acquired, and light source information that corresponds to the acquired position is acquired. Color information of a region including the object in the captured image is corrected based on the acquired light source information.

Abstract: A system and method for method for identifying key points in a multimedia data element (MMDE). The method includes: identifying, via a computer vision system, a plurality of candidate key points in the MMDE, wherein a size of each candidate key point is equal to a predetermined size, wherein a scale of each candidate key point is equal to a predetermined scale; analyzing the plurality of candidate key points to determine a set of properties for each candidate key point; comparing the sets of properties of the plurality of candidate key points; and selecting, based on the comparison, a plurality of key points from among the candidate key points.

Abstract: A feature determination apparatus and method adapted to multiple object sizes are provided. The apparatus individually supplies each of the object images to a convolution neural network having several convolution layers to generate multiple feature maps corresponding to each object image. The apparatus calculates a feature amount of each feature image of each object image. The apparatus determines an invalid layer start number of each object image according to a preset threshold and the feature amount corresponding to each object image. The apparatus determines a feature map extraction recommendation for each of a plurality of object sizes according to a size of each object image and the invalid layer start number of each object image.

Abstract: The present application discloses a method and apparatus for obtaining an image of a form sheet, which belongs to the field of image processing technology. The method includes: obtaining graphic code information from an image to be processed, wherein the graphic code information comprises at least position information of a graphic code; determining form sheet image feature information in the image to be processed based on the position information of the graphic code; and obtaining an image of a form sheet from the image to be processed based on the form sheet image feature information. In the present application, there is no manual operation of obtaining the image of the form sheet is needed, and thus the resources are saved and the efficiency of obtaining the image of the form sheet is improved.