Abstract: The disclosed technology is generally directed to optical text recognition for forms. In one example of the technology, line grouping rules are generated based on the generic forms and a ground truth for the generic forms. Line groupings are applied to the generic forms based on the line grouping rules. Feature extraction rules are generated. Features are extracted from the generic forms based on the feature extraction rules. A key-value classifier model is generated, such that the key-value classifier model is configured to determine, for each line of a form: a probability that the line is a value, and a probability that the line is a key. A key-value pairing model is generated, such that the key-value pairing model is configured to predict, for each key in a form, which value in the form corresponds to the key.

Abstract: An image processing device mounted to a vehicle includes a motion feature extracting unit is configured to create motion information indicative of a motion of an observation target from a captured image including the observation target around the vehicle, a distance feature extracting unit is configured to calculate a distance from the vehicle to the observation target based on the captured image, and a recognition unit is configured to execute a recognition process of the observation target based on process results of the motion feature extracting unit and the distance feature extracting unit. The motion feature extracting unit and the distance feature extracting unit have respective operation processing amounts changed based on a predetermined vehicle state.

Abstract: An illustrative, non-limiting method for resolving Region-On-Interest (ROI) overlaps includes receiving Simultaneous Localization and Mapping (SLAM) data obtained by a plurality of co-located Head-Mounted Devices (HMD), where the SLAM data comprises a plurality of landmarks and each HMD has an Information Handling System (IHS) in communication therewith. Each IHS may produce a subset of the plurality of landmarks based upon a respective HMD's SLAM data, and each IHS is: local with respect to the plurality of HMDs, at an edge of a network serving an HMD, or on a cloud. The method may include identifying an overlap between a first landmark produced by a first IHS and a second landmark produced by a second IHS, selecting between the first and second landmarks, at least in part, based upon the locations of the first and second IHSs, and rendering a map for display by a given HMD using the selected landmark.

Abstract: An ultra-dense electrode-based brain imaging system with high spatial and temporal resolution. A Sparsity and Smoothness enhanced Method Of Optimized electrical TomograpHy (s-SMOOTH) based reconstruction technique to improve the spatial resolution and localization accuracy of reconstructed brain images is described. Also described is a graph Fractional-Order Total Variation (gFOTV) based reconstruction technique to improve the spatial resolution and localization accuracy of reconstructed brain images.

Abstract: Disclosed is a deep learning-based object recognition system which includes a data collection process executor configured to perform a data collection process of collecting values of the PIR sensor according to a sampling period; a data classification process executor configured to perform a data classification process of inputting the collected values of the PIR sensor to a model of an artificial neural network and transmitting a result of the inputting the collected values to a cloud using a RESTful API; and an object recognition cloud system configured to store information received from the data classification process executor in a database, transmit the information when a web application requests the information, and represent information collected and classified by devices using the RESTful API.

Abstract: According to some aspects, a system for identifying a printed object is provided. The system may include an imaging device configured to image a three-dimensional (3D) printed object and a processor coupled to the imaging device. The processor may be configured to receive information generated as a result of imaging the 3D printed object with the imaging device, identify a set of features for the 3D printed object using the received information, compare the set of features with multiple reference sets of features, each of the multiple reference sets of features being associated with a respective 3D object in a plurality of 3D objects, and identify, based at least in part on results of the comparison, a 3D object from the plurality of 3D objects that matches the 3D printed object.

Abstract: Disclosed herein are system, method, and computer program product embodiments for verifying a document. An embodiment operates by receiving an initial image of a document and identifying a first imperfection on a feature of the document. Thereafter, the feature of the initial image is segmented into a first and second partially overlapping segment corresponding to a first and second portion of the feature, respectively. Subsequently, the feature's first portion is determined to be free from the first imperfection, and the first imperfection is identified in the feature's second segment. After receiving a second image of the document including the feature's second portion, the feature's second portion in the second image is determined to be free from a second imperfection. As a result, an image of the feature is created based on the first and second portions of the feature derived from the initial and second images, respectively.

Abstract: An example apparatus for correcting gaze in images includes an image receiver to receive an image comprising an eye and a target angle set to a center. The apparatus also includes a bidirectionally trained convolutional neural network (CNN) to receive the image and the target angle from the image receiver and generate a vector field and a brightness map based on the image and the target angle. The apparatus further includes an image corrector to generate a gaze corrected image based on the vector field and the brightness map.

Abstract: In one embodiment, a feature map of an image having h×w pixels and a patch having one or more pixels of the image are received. The patch has been processed by a first set of layers of a convolutional neural network and contains an object centered within the patch. The patch is then processed using the feature map and one or more pixel classifiers of a classification layer of a deep-learning model, where the classification layer includes h×w pixel classifiers, with each pixel classifier corresponding to a respective pixel of the patch. Each of the pixel classifiers used to process the patch outputs a respective value indicating whether the corresponding pixel belongs to the object centered in the patch.

Abstract: A method, system and device for determining the properties of hair and skin. The method, system and device facilitates the selection and application of an application of a product to the hair or skin of an individual to achieve a desired target result. The hair or skin properties measured may include the hair composition and chemical components of the hair, as well as substances that may coat the hair, and other substances on, absorbed or adsorbed on or into the hair. A hyper-spectral imaging component is employed to obtain information from the hair or skin and that information is used to provide a produce used to produce the target result.

Abstract: There are provided systems and methods for performing automated content evaluation. In one implementation, the system includes a hardware processor and a system memory storing a software code including a predictive model trained based on an audience response to training content. The hardware processor executes the software code to receive images, each image including facial landmarks of an audience member viewing the content during its duration, and for each image, transforms the facial landmarks to a lower dimensional facial representation, resulting in multiple lower dimensional facial representations of each audience member. For each of a subset of the lower dimensional facial representations of each audience member, the software code utilizes the predictive model to predict one or more responses to the content, resulting in multiple predictions for each audience member, and classifies one or more time segment(s) in the duration of the content based on an aggregate of the predictions.

Abstract: The present invention discloses a method operable on a digital electronic device comprising constrained processing unit employing a limited computer-readable storage medium, also known as a digital memory, for allowing an object classification process to be executed on an image. The object classification process may be allowed by a processing unit interlocking with a digital memory unit, which receives an image representing a digital image captured by light incident on an image sensor, denoted herein as an original image. In some cases, a computerized process operable on the digital processing may identify a list of pixel arrays located at the original image, and thereby allow a classification process to be operated on these pixel arrays. In some cases, a process operable on the digital processing may grant access to another computerized process to perform the classification process.

Abstract: An embodiment of the invention may include a method, computer program product and computer system for image identification and classification. The method, computer program product and computer system may include a computing device which may receive one or more images of a first object from at least two angles linguistic data associated with the first object. The computing device may input the one or more images of the first object into one or more first neural networks and the linguistic data of the first object into one or more second neural networks. The computing device may combine the output of the one or more first neural networks and the one or more second neural networks and generate an identification model based on the combined output of the one or more first neural networks and the one or more second neural networks.

Abstract: A system and method for capturing an image of a physical space and designing a multidimensional artificial depiction of design option modifications thereto, using a mobile device or computer, using AR. The system includes CR software programs to identify and segment surfaces and objects in the physical and multidimensional artificial spaces; to determine lighting in the spaces, and to map locations into 3D depictions. Also included is a user interface to provide user input and computer feedback, a user control component to change at least two features within the depiction thus generating an overlay of the multidimensional artificial space as to the physical space.

Abstract: Disclosed herein are systems, methods, and non-transitory computer-readable storage media for performing speaker verification. A system configured to practice the method receives a request to verify a speaker, generates a text challenge that is unique to the request, and, in response to the request, prompts the speaker to utter the text challenge. Then the system records a dynamic image feature of the speaker as the speaker utters the text challenge, and performs speaker verification based on the dynamic image feature and the text challenge. Recording the dynamic image feature of the speaker can include recording video of the speaker while speaking the text challenge. The dynamic feature can include a movement pattern of head, lips, mouth, eyes, and/or eyebrows of the speaker. The dynamic image feature can relate to phonetic content of the speaker speaking the challenge, speech prosody, and the speaker's facial expression responding to content of the challenge.

Abstract: Methods for white-balancing images are provided. These methods include determining, for an input image, a chrominance histogram for the pixels of the input image. The determined histogram is a toroidal chrominance histogram, with an underlying, toroidal chrominance space that corresponds to a wrapped version, of a standard flat chrominance space. The toroidal chrominance histogram is- then convolved with a fitter to generate a two-dimensional heat map that is then used to determine art estimated chrominance of i|lummaiioB present id the input image; This can include fitting a bivariate von Mises distribution, or some other circular and/or toroidal, probability distribution, to the determined two-dimensional heat map. These methods for estimating illumination chrominance values for input images have reduced computational costs and increased speed relative to other methods for determining image illuminant chrominance values.

Abstract: The present disclosure relates to systems and methods for automated identification of presence of a facial action from sequential images. The systems and methods may obtain a plurality of sequential candidate image frames containing a facial object. Each of the plurality of candidate image frames may include a plurality of feature points associated with an eye of the facial object. For each of the plurality of sequential candidate image frames, the systems and methods may determine an area of a polygon based on the plurality of feature points. The systems and methods may determine a distance associated with the eye of the facial object. The systems and methods may determine an action parameter based on the area and the distance. The systems and methods may identify the presence of a blink action in response to that the action parameters satisfy one or more preset conditions.

Abstract: Cameras capture time-stamped images of predefined areas. At least one image includes a representation of a portion of an individual with other portions of the individual occluded within the image. Pixel attributes for the portion of the individual are identified and provided as a box or set of coordinates for tracking the individual within the image and in subsequent images taken.

Abstract: The invention relates to a method for imaging a three-dimensional object to be examined. According to said method, a three-dimensional parameterized area is determined which is in conformity with an anatomic structure of the three-dimensional object to be examined. The three-dimensional parameterized area is imaged onto a two-dimensional parameterized area. The three-dimensional object to be examined is represented by imaging pixels that are associated with the three-dimensional parameterized area onto the two-dimensional parameterized area. The invention further relates to a method for determining a camera position in a three-dimensional image recording of an object to be examined. The invention also relates to a method for representing a section of an object to be examined. The invention finally relates to a device for imaging a three-dimensional object to be examined.

Abstract: According to some aspects, an information processing apparatus is provided. The information processing apparatus includes at least one processor configured to receive an image of a plurality of images. The information processing apparatus further includes at least one storage medium configured to store processor-executable instructions that, when executed by the at least one processor, perform a method. The method includes setting at least one axial direction in the image, wherein the image includes an analysis target. The method further includes determining motion information for the analysis target by analyzing motion of the analysis target to identify, with respect to the at least one axial direction, at least one of a motion amount of the analysis target and a motion direction of the analysis target.