Abstract: A parasite analysis system includes a pressure vessel configured to store a biological sample, an imaging cell connected to the pressure vessel, and a waste depository connected to the imaging cell. An input valve controls whether biological sample can flow from the pressure vessel into the imaging cell and an output valve controls whether biological sample can flow from the imaging cell into the waste depository. The parasite analysis system also includes a camera that captures a chronological set of images of a portion of the biological sample in the imaging cell and an image analysis system that analyzes the chronological set of images to generate an estimate of a number of parasites in the portion of the biological sample. Estimates for multiple portions of the biological sample may be generated and sampling techniques used to estimate the number of parasites in the entire biological sample.

Abstract: A method for guiding a robot equipped with a camera to facilitate three-dimensional (3D) reconstruction through sampling based planning includes recognizing and localizing an object in a two-dimensional (2D) image. The method also includes computing 3D depth maps for the localized object. A 3D object map is constructed from the depth maps. A sampling based structure is grown around the 3D object map and a cost is assigned to each edge of the sampling based structure. The sampling based structure may be searched to determine a lowest cost sequence of edges that may, in turn be used to guide the robot.

Abstract: An information processing apparatus comprises an image generation unit configured to generate, based on a first image in which a transparent object having transparency is captured and a second image in which a target object is captured, a reproduced image in which the target object which is at least partially covered by the transparent object is reproduced; and a creation unit configured to create, based on the reproduced image, a model for recognizing the target object which is at least partially covered by the transparent object.

Abstract: To determine real-world information about objects moving in a scene, the camera capturing the scene is typically calibrated to the scene. Automatic scene calibration can be accomplished using people that are found moving about in the scene. During a calibration period, a video content analysis system processing video frames from a camera can identify blobs that are associated with people. Using an estimated height of a typical person, the video content analysis system can use the location of the person's head and feet to determine a mapping between the person's location in the 2-D video frame and the person's location in the 3-D real world. This mapping can be used to determine a cost for estimated extrinsic parameters for the camera. Using a hierarchical global estimation mechanism, the video content analysis system can determine the estimated extrinsic parameters with the lowest cost.

Abstract: A system and method are provided for capturing an image with correct skin tone exposure. In use, one or more faces are detected having threshold skin tone within a scene. Next, based on the detected one or more faces, the scene is segmented into one or more face regions and one or more non-face regions. Next, one or more images of the scene are captured. Further, in response to the capture, the one or more face regions are processed to generate a final image.

Abstract: A face detection apparatus detecting a face from an image which is captured by an imaging unit and includes the face, includes: a position detection unit that detects a position of a face part of the face from the image; an initial state determination unit that determines a model initial state on the basis of the position of the face part; and a model fitting unit that generates a model of the face on the basis of the image by using the model initial state.

Abstract: This invention provides a method of face recognition and a device for implementing said method, the method comprising the steps of: (a) performing a face recognition algorithm on an image of a user to determine an identity of the user, the image having an illumination level; (b) determining if a reliability measurement of the identity of the user determined by the face recognition algorithm meets a threshold indicating reliable identification and, if the reliability measurement is below the threshold indicating reliable identification; (c) receiving a new image of the user, the new image having an adjusted illumination level; and (d) performing a further face recognition algorithm on the new image of the user to determine the identity of the user.

Abstract: A three-dimensional (3D) image recognition system includes a first imaging sensor capable of collecting a first wavelength range of light and a second imaging sensor capable of collecting a second wavelength range of light. The first imaging sensor and the second imaging sensor are placed apart. The 3D image recognition system also includes a processor configured to identify at least one landmark area of a first image of an object collected by the first imaging sensor, and identify at least one matching landmark area in a second image of the object collected by the second imaging sensor. The processor is further configured to extract the 3D information of the object from the at least one landmark area of the images collected.

Abstract: A method of predicting the probability of injuries in ice hockey which consists of analysing data from accelerometers and gyroscopes on the torso of a player and assessing the work load of the muscles of the player as an indicator of the probability of injury. One aspect provides a method of predicting the probability of groin injuries in ice hockey which consists of analysing data from accelerometers and gyroscopes on the torso of a player and assessing groin load as an indicator of the probability of groin injury. A second aspect provides method of managing player availability by limiting injuries which consists of analysing data from accelerometers and gyroscopes on the torso of a player and identifying and counting the number of slap shots executed as a means of assessing the player load as an indicator of probability of injury.

Abstract: An object detection system for a marine vessel having at least one marine drive includes at least one image sensor positioned on the marine vessel and configured to capture an image of a marine environment on or around the marine vessel, and a processor. The object detection system further includes an image scanning module executable on the processor that receives the image as input. The image scanning module includes an artificial neural network trained to detect patterns within the image of the marine environment associated with one or more predefined objects, and to output detection information regarding a presence or absence of the one or more predefined objects within the image of the marine environment.

Abstract: Many conventional video processing algorithms attempting to detect human presence in a video stream often generate false positives on non-human movements such as plants moving in the wind, rotating fan, etc. To reduce false positives, a technique exploiting temporal correlation of non-human movements can accurately detect human occupancy while reject non-human movements. Specifically, the technique involves performing temporal analysis on a time-series signal generated based on an accumulation of foreground maps and an accumulation of motion map and analyzing the running mean and the running variance of the time-series signal. By determining whether the time-series signal is correlated in time, the technique is able to distinguish human movements and non-human movements. Besides having superior accuracy, the technique lends itself to an efficient algorithm which can be implemented on low cost, low power digital signal processor or other suitable hardware.

Abstract: The present invention provides a method using a trained and activated artificial intelligence entity for detecting in photo images, moving or still, the identity of target people who are present or former members of the U.S. armed services, or are related thereto, from sources available through the Internet or other wide area computer network.

Abstract: A method for performing image processing of a physical document image. The method may include receiving an image file including an image of a physical document and a control target. The method may include measuring control target attributes of the control target. Measuring the control target attributes of the control target may include numerically measuring a degree of focus of the image file based on the control target. The method may include performing a static crop of the image file based on the control target attributes by removing the control target from the image file to create a cropped image.

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: Determining if a document is a text page includes partitioning the document into a plurality of cells, scaling each of the cells to a standardized number of pixels to provide a corresponding snippet for each of the cells, using a classifier to examine the snippets to determine which of the cells are classified as text and which of the cells are not classified as text, determining a volume of text for the document based on a total amount of text in the document corresponding to a sum of an amount of text in each of the cells classified as text, and determining that the document is a text page in response to the total amount exceeding a pre-determined threshold. In response to the total amount being less than the pre-determined threshold, cells not classified as text may be examined further. The classifier may be provided by training a neural net.

Abstract: The methods and apparatus for repetitive iris recognition include an apparatus for repetitively imaging an iris. The apparatus comprises a housing defining a substantially planar outer surface, a first axis normal to the substantially planar outer surface, and a curved outer surface, opposite the substantially planar outer surface, defined by a curve rotated about the first axis. The apparatus further comprises an infrared sensor array, disposed within the housing, to acquire an infrared image of an iris via at least one first aperture in the substantially planar outer surface. The apparatus further comprises at least one infrared illuminator, disposed within the housing, to illuminate the iris during acquisition of the infrared image of the iris via at least one second aperture in the substantially planar outer surface.

Abstract: A system and various methods for processing an image to produce a hierarchical image representation model, segment the image model using shape criteria to produce positive and negative training data sets as well as a search-space data set comprising shapes matched to a search query provided as input, and using the training data sets to train a machine learning model to improve recognition of shapes that are similar to an input query without being exact matches, to improve object recognition.

Abstract: A system for simplified generation of systems for analysis of satellite images to geolocate one or more objects of interest. A plurality of training images labeled for a study object or objects with irrelevant features loaded into a preexisting feature identification subsystem causes automated generation of models for the study object. This model is used to parameterize pre-engineered machine learning elements that are running a preprogrammed machine learning protocol. Training images with the study are used to train object recognition filters. This filter is used to identify the study object in unanalyzed images. The system reports results in a requestor's preferred format.

Abstract: A method and device for detecting an overhead cable from an aerial vessel, for instance from a helicopter in motion, and specifically for use in a landing preparing procedure. The method includes providing, from a camera arranged in the aerial vessel, a plurality of images of an area beneath the aerial vessel when the aerial vessel is in motion; detecting lines in the images; determining if the lines represent lines at ground level or lines at a level above ground level; and determining if a line at a level above ground level represents a possible overhead cable. The step of detecting lines in the images may include a modified SUSAN algorithm. The step of determining if the lines represent lines at ground level or lines above ground level may include parallax calculation.

Abstract: A system for observing agricultural samples includes a chassis suspended on an elevated cable or rail, an actuator disposed within the chassis for moving the chassis forward and backward along the elevated cable or rail, a camera mounted on or within the chassis and configured to acquire image data of an area below the elevated cable or rail including an agricultural sample, and a processor disposed within the chassis for receiving image data from the camera, autonomously controlling the actuator to move the chassis along the elevated cable or rail, and assessing a condition of the agricultural sample from the received image data.

Abstract: A wearable imaging apparatus is provided for capturing and processing images from an environment of a user. In one implementation, the wearable apparatus may be configured with a memory for storing privacy mode triggers and associated automatically variable privacy mode settings, and at least one processing device. The processing device may analyze the images captured by the wearable apparatus, and recognize the presence of at least one of the privacy mode triggers within the images. After recognizing the at least one trigger, the processing device may cause one or more adjustments to the wearable apparatus based on the privacy mode settings associated with the at least one trigger.

Abstract: A control apparatus includes a detection unit, an association unit, and an output control unit. The detection unit detects a person from an image which includes a plurality of persons. The association unit associates the persons included in the image with each other based on at least one of a position of the person detected by the detection unit, directions of faces of the persons included in the image, and distances between the persons included in the image. The output control unit causes an output unit to output information that is indicative of a relation of the detected person with respect to other persons included in the image based on a result of association performed by the association unit.

Abstract: A method for temperature control includes acquiring image data, and analyzing the image data to identify one or more individuals within the image data. A level of thermal comfort is estimated for each of the identified individuals based on the image data. The determined level of thermal comfort indicates whether each individual appears to be feeling too hot, appears to be feeling too cold or appears to be feeling satisfied with an ambient temperature. An air-conditioning unit or a heating unit is controlled based on the estimation of the level of thermal comfort for each of the identified individuals.

Abstract: Systems, methods, and devices for curating audiovisual content are provided. A mobile image capture device can be operable to capture one or more images; receive an audio signal; analyze at least a portion of the audio signal with a first machine-learned model to determine a first audio classifier label descriptive of an audio event; identify a first image associated with the first audio classifier label; analyze the first image with a second machine-learned model to determine a desirability of a scene depicted by the first image; and determine, based at least in part on the desirability of the scene depicted by the first image, whether to store a copy of the first image associated with the first audio classifier label in the non-volatile memory of the mobile image capture device or to discard the first image without storing a copy of the first image.

Abstract: The invention provides a method of analyzing a sporting activity comprising receiving location data associated with a first player, a second player, and a game object; detecting, by a processing device, a game event associated with the first player; determining, in view of the received location data, a plurality of relative locations of the first player at a time prior to the game event, wherein the plurality of relative locations comprise a relative location in relation to the second player and a relative location in relation to a playing area; and classifying the game event based on the plurality of relative locations of the first player at the time prior to the game event.

Abstract: Processing of a continuous, high-resolution video data stream for reduced data storage is provided. The processing includes: analyzing the continuous, high-resolution data stream for one or more specified image elements of interest; identifying at least one image element of the one or more specified image elements of interest in the continuous video data stream; passing one or more video segments of the continuous video data stream as high-resolution video, the one or more video segments containing the identified at least one image element of the one or more specified image elements of interest; and reducing resolution of other portions of the continuous video data stream than the one or more video segments to low-resolution video, wherein low-resolution video includes less data than high-resolution video.

Abstract: A method of selecting at least one video frame of a video sequence. A plurality of faces is detected in at least one video frame of the video sequence. An orientation of the detected faces is tracked over a series of subsequent video frames to determine whether a first detected face is turning towards a second detected face. The method then determines, using the tracked orientation of the detected faces, a portion of the video sequence in which the first and second detected faces are oriented towards each other for at least a predetermined number of frames defining a gaze fixation of the detected faces. At least one video frame is selected from the determined portion of the video sequence, the selected video frame capturing the gaze fixation of the detected faces.

Abstract: A device for previewing a video includes a memory device configured to store instructions, and one or more processors configured to execute the instructions to receive a video captured by a camera and analyze the video. The one or more processors are also configured to detect a special event from the analysis of the video. The special event includes at least one of an object, a moving object, or a sound detected in the video. The one or more processors are further configured to extract, from the video, at least one of a video preview frame or audio segment corresponding to the special event, and generate a preview video including the at least one of the video preview frame or audio segment corresponding to the special event.

Abstract: A vehicle controller receives images from a camera upon arrival and upon departure. A location of the vehicle may be tracked and images captured by the camera may be tagged with a location. A departure image may be compared to an arrival image captured closest to the same location as the arrival image. A residual image based on a difference between the arrival and departure images is evaluated for anomalies. Attributes of the anomaly such as texture, color, and the like are determined and the anomaly is classified based on the attributes. If the classification indicates an automotive fluid, then an alert is generated.

Abstract: Examples disclosed herein relate to a person moving in a physical space. In one aspect, a method is disclosed. The method may include obtaining at least two images of a person from at least two cameras directed at a physical space, where the physical space may include a plurality of designated areas. The method may also include obtaining metadata associated with the images, based on the images and the metadata determining within the plurality of designated areas a set of designated areas visited by the person, for each designated area within the set of designated areas, determining an area information, and updating a database based on the set of designated areas and based on at least a portion of the area information associated with each designated area within the set of designate areas.

Abstract: A checkout lane management system is described that uses object recognition to detect whether one or more items is under a shopping cart during a customer checkout process. The checkout lane management system may comprise one or more of cameras for collecting a stream of images focused on a checkout lane. The checkout lane management system may also comprise one or more of a display device and a speaker device as part of a checkout lane device for displaying a BOB indicator or sounding the BOB indicator.

Abstract: A plurality of pairs of video cameras and interrogation devices may be placed in a public place along various paths that a person-of-interest might be expected to move. The person-of-interest is then located in multiple images acquired, collectively, by multiple video cameras. From each of the interrogation devices that are paired with these video cameras, a subset of the captured identifiers is obtained. Candidate identifiers are then restricted to those identifiers that are included in each of the subsets. A given identifier may be rejected as a candidate identifier. To automatically locate the person-of-interest in the images acquired by the “paired” video cameras, a processor may utilize video-tracking techniques to automatically track the person-of-interest, such that the person-of-interest is not “lost.” By virtue of utilizing such tracking techniques, the person-of-interest may be repeatedly located automatically, and with minimal chance of a false detection.

Abstract: A method of identifying a condition of a road surface includes capturing at least a first image of the road surface with a first camera, and a second image of the road surface with a second camera. The first image and the second image are tiled together to form a combined tile image. A feature vector is extracted from the combined tile image using a convolutional neural network, and a condition of the road surface is determined from the feature vector using a classifier.

Abstract: An approach is provided for parametric representation of lane lines. The approach involves segmenting an input image into grid cells. The approach also involves processing a portion of the input image in each grid cell to detect lane lines. The approach further involves, for each grid cell in which lane lines are detected, determining intercepts of the lane lines with edges of the grid cell, and slopes of the lane lines at the intercepts. The approach further involves generating a parametric representation of the lane lines for each grid cell. The parametric representation encodes the intercepts and slopes into a data structure for each grid cell. The approach further involves providing an output parametric representation for the input image that aggregates the parametric representations of each grid cell.

Abstract: A method of lane detection for a non-transitory computer readable storage medium storing one or more programs is disclosed. The one or more programs include instructions, which when executed by a computing device, cause the computing device to perform the following steps comprising: generating a ground truth; off-line training a lane detection algorithm by using the ground truth, the lane detection algorithm using parameters that express a lane marking in an arc; on-line generating a predicted lane marking; comparing the predicted lane marking against the ground truth; and off-line refining the lane detection algorithm.

Abstract: A method for detecting lane elements, which are unit regions including pixels of lanes in an input image, to plan the drive path of an autonomous vehicle by using a horizontal filter mask is provided. The method includes steps of: a computing device acquiring a segmentation score map from a CNN using the input image; instructing a post-processing module, capable of performing data processing at an output end of the CNN, to generate a magnitude map by using the segmentation score map and the horizontal filter mask; instructing the post-processing module to determine each of lane element candidates per each of rows of the segmentation score map by referring to values of the magnitude map; and instructing the post-processing module to apply estimation operations to each of the lane element candidates per each of the rows, to thereby detect each of the lane elements.

Abstract: A driver monitoring apparatus includes: a driver monitoring camera configured to obtain an image of a user while adjusting an irradiation range of irradiated light; and a driver state sensor configured to request the driver monitoring camera to adjust the irradiation range depending on an operation mode. The driver monitoring apparatus is capable of improving reliability in monitoring a driver state by controlling an irradiation range and intensity of lighting depending on an operation mode and an external environment at the time of monitoring the driver state to irradiate light to only a region of interest.

Abstract: In general, embodiments of the technology relate to extracting content from documents. More specifically, embodiments of the technology relate to using fuzzy regular expressions to process content results obtained from one or more documents in order to extract content for these documents. Further, embodiments of the technology enable the format modification of the content after the content has been identified and extracted from the documents.

Abstract: Examples of a system and method for identifying double feed documents are generally described. A method may include scanning a plurality of documents using a scanner in a first configuration and rejecting identified potential double feeds of checks from the plurality of checks in the first configuration. The method may include rescanning identified potential double feeds in a second configuration.

Abstract: Exemplary embodiments are directed to biometric analysis systems including one or more illumination sources configured to provide dim illumination to a scene including an object and configured to provide flash illumination to the object in the scene. The biometric analysis systems include a rolling shutter camera configured to capture one or more images. The biometric analysis systems include an adaptive trigger module configured to analyze the scene to detect the object in the scene during dim illumination of the scene, determine a position in a frame of the rolling shutter camera that coincides with the detected object in the scene, and arrange a delay between a start of image writing by the rolling shutter camera and a trigger of the one or more illumination sources such that a stripe of the flash illumination coincides with the detected object in the scene.

Abstract: A character input method using a touch screen, in which one or more areas requiring user input is defined in the touch screen, pre-recognized information is defined for each of the defined areas, character information is received by a user in one or more user desired areas among the defined areas, the character information is recognized using a character recognizer, and the recognized character information is updated in the user desired areas.

Abstract: A method of defining a path model from a set of realistic paths is provided, where each path in the set of realistic paths is expanded on piece-wise polynomial basis, and a respective centroid function and sequence of eigenfunctions calculated for each expanded representation. A set of principle paths representing the major variation of this set of paths is obtained describing the variations of the set of realistic paths with respect to the centroid. The path model thus comprises a linear combination of said principle paths. The path model may be used as the basis for the generation of new curves having similar characteristics to the original set of realistic paths.

Abstract: A system for determining a viable area automatically and accurately includes an interface and a processor. The interface is configured to receive location information. The processor is configured to determine multi-view imagery data based on the location information, to determine a point cloud using the multi-view imagery data, determine one or more features using the point cloud, determine viable area using the one or more features, and determine a solar irradiance of the viable area.

Abstract: Methods, apparatuses, and embodiments related to analyzing the content of digital images. A computer extracts multiple sets of visual features, which can be keypoints, based on an image of a selected object. Each of the multiple sets of visual features is extracted by a different visual feature extractor. The computer further extracts a visual word count vector based on the image of the selected object. An image query is executed based on the extracted visual features and the extracted visual word count vector to identify one or more candidate template objects of which the selected object may be an instance. When multiple candidate template objects are identified, a matching algorithm compares the selected object with the candidate template objects to determine a particular candidate template of which the selected object is an instance.

Abstract: Systems and methods for identifying an object and presenting additional information about the identified object are provided. The techniques of the present invention can allow the user to specify modes to help with identifying objects. Furthermore, the additional information can be provided with different levels of detail depending on user selection. Apparatus for presenting a user with a log of the identified objects is also provided. The user can customize the log by, for example, creating a multi-media album.

Abstract: It is provided an object detection method to be performed by an image analysis device configured with a computer including a processor and a memory, the object detection method comprising: a first step of receiving, by the processor, input of an image and storing the image to the memory; a second step of extracting, by the processor, a partial area which is a part of the input image and storing the partial area to the memory; a third step of evaluating, by the processor, symmetry of the partial area using features extracted from the partial area and reflected features obtained by applying reflection to the features; and a fourth step of determining, by the processor, that the partial area is an object area including an object to be detected in a case where the symmetry of the partial area is evaluated as higher than a predetermined threshold.

Abstract: Provided are a system and method of detecting moving objects. The system stores pixel information regarding each of pixels included in frames of the video in a storage, sets a background model comprising at least one background model element, the at least one background model element indicating at least one of a reference visual property and a reference usage time, determines whether the pixels are background pixels or foreground pixels by comparing the pixel information with the at least one background model element, and updating the background model based on a result of the comparing.

Abstract: A system and method for recognizing a medication are provided. The method includes the steps of presenting a medication to a medication identification apparatus, the medication identification apparatus adapted to visually image the presented medication and visually imaging the presented medication. A predetermined transformation may be applied to the visually imaged presented medication, the predetermined transformation adjusting one or more parameters of the visually imaged presented medication along one or more dimensions. Finally, the transformed visually imaged medication is compared to a medication library to determine a closest match therebetween.

Abstract: A method can include comparing a first feature vector detailing features of an image of a newsfeed of a user of users of a social network to a subset of second feature vectors detailing features of newsfeeds presented to the users of the social network; and in response to determining the first feature vector matches a second feature vector of the subset of second feature vectors, providing a name, profile data, and profile picture of a user associated with the newsfeed.

Abstract: The invention is related to a method of determining a similarity transformation between first coordinates and second coordinates of 3D features, comprising providing a first plurality of 3D features having first coordinates in a first coordinate system which is associated with a first geometrical model of a first real object, wherein the first plurality of 3D features describes physical 3D features of the first real object, providing a second coordinate system, providing image information associated with a plurality of images captured by at least one camera, for each respective 3D feature of at least part of the first plurality of 3D features, wherein the respective 3D feature is captured by at least two of the plurality of images, determining camera poses of the at least one camera in the second coordinate system while the at least two of the plurality of images are captured, determining for the respective 3D feature a second coordinate in the second coordinate system according to the at least two of the plu