Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for classifying videos using neural networks. One of the methods includes obtaining a temporal sequence of video frames, wherein the temporal sequence comprises a respective video frame from a particular video at each of a plurality time steps; for each time step of the plurality of time steps: processing the video frame at the time step using a convolutional neural network to generate features of the video frame; and processing the features of the video frame using an LSTM neural network to generate a set of label scores for the time step and classifying the video as relating to one or more of the topics represented by labels in the set of labels from the label scores for each of the plurality of time steps.

Abstract: A soft back-up pin state checking device including a soft back-up pin formed of a bottom-side holder portion, and a top-side soft pin portion which is held by the holder portion in an exchangeable manner. A ring-shaped background portion which is possible to be distinguished from the soft pin portion and be recognized using image processing is provided on a top end surface of a cylindrical fitting portion which is outside of an outer circumference of the soft pin portion in the holder portion of the soft back-up pin.

Abstract: A method for enhancing user liveness detection is provided that includes calculating, by a computing device, a first angle and a second angle for each frame in a video of captured face biometric data. The first angle is between a plane defined by a front face of the terminal device and a vertical axis, and the second angle is between the plane defined by the front face of the terminal device and a plane defined by the face of the user. Moreover, the method includes creating a first signal from the first angles and a second signal from the second angles, calculating a similarity score between the first and second signals, and determining the user is live when the similarity score is at least equal to a threshold score.

Abstract: A method of aiding the diagnosis of otitis media in a patient includes obtaining image data in a processor apparatus of a computing device, the image data being associated with at least one electronic image of a tympanic membrane of the patient, calculating, a plurality of image features, each image feature being calculated based on at least a portion of the image data, classifying the at least one electronic image as a particular type of otitis media using the plurality of image features, and outputting an indication of the particular type of otitis media. Also, a system for implementing such a method that includes an output device and a computing device.

Abstract: A detection method executed by a computer, the detection method includes detecting a plurality of pupil candidates from a face image region in an image of a subject based on specific shape information, and identifying at least one pupil candidate as a pupil from among the plurality of pupil candidates based on brightness information related to an image region outside of the face image region and learning information indicating a relationship between the brightness information and a size of the pupil.

Abstract: Computer program products for discriminating hand and machine print from each other, and from signatures, are disclosed and include program code readable and/or executable by a processor to: receive an image, determine a color depth of the image; reducing the color depth of non-bi-tonal images to generate a bi-tonal representation of the image; identify a set of one or more graphical line candidates in either the bi-tonal image or the bi-tonal representation, the graphical line candidates including true graphical lines and/or false positives; discriminate any of the true graphical lines from any of the false positives; remove the true graphical lines from the bi-tonal image or the bi-tonal representation without removing the false positives to generate a component map comprising connected components and excluding graphical lines; identify one or more of the connected components in the component map; and output and/or display and indicator of each of the connected components.

Abstract: An image processing apparatus 1 includes a feature point set acquirer, a degree-of-scattering acquirer, a fundamental matrix acquirer and a processor. The feature point set acquirer acquires feature point sets, each including a plurality of feature points within an image. The degree-of-scattering acquirer acquires a degree of scattering indicating a degree of scattered distribution of the feature points. The fundamental matrix acquirer acquires a fundamental matrix, based on coordinates of the feature points p included in feature point sets acquired by the feature point set acquirer and the degree-of-scattering acquired by the degree-of-scattering acquirer, acquires a fundamental matrix indicating the epipolar geometric relationship between the image and another image in a video including the image. The processor performs image processing of the video based on the fundamental matrix acquired by the fundamental matrix acquirer.

Abstract: Methods and systems for determining coordinates for an area of interest on a specimen are provided. One system includes one or more computer subsystems configured for, for an area of interest on a specimen being inspected, identifying one or more targets located closest to the area of interest. The computer subsystem(s) are also configured for aligning one or more images for the one or more targets to a reference for the specimen. The image(s) for the target(s) and an image for the area of interest are acquired by an inspection subsystem during inspection of the specimen. The computer subsystem(s) are further configured for determining an offset between the image(s) for the target(s) and the reference based on results of the aligning and determining modified coordinates of the area of interest based on the offset and coordinates of the area of interest reported by the inspection subsystem.

Abstract: Cone Beam Breast CT, (CBBCT) is a three-dimensional breast imaging modality with high soft tissue contrast, high spatial resolution and no tissue overlap. CBBCT-based computer aided diagnosis (CBBCT-CAD) technology is a clinically useful tool for breast cancer detection and diagnosis that will help radiologists to make more efficient and accurate decisions. The CBBCT-CAD is able to: 1) use 3D algorithms for image artifact correction, mass and calcification detection and characterization, duct imaging and segmentation, vessel imaging and segmentation, and breast density measurement, 2) present composite information of the breast including mass and calcifications, duct structure, vascular structure and breast density to the radiologists to aid them in determining the probability of malignancy of a breast lesion.

Abstract: Apparatuses, methods and storage medium associated with processing an image are disclosed herein. In embodiments, a method for processing one or more images may include generating a plurality of pairs of keypoint features for a pair of images. Each pair of keypoint features may include a keypoint feature from each image. Further, for each pair of keypoint features, corresponding adjoin features may be generated. Additionally, for each pair of keypoint features, whether the adjoin features are similar may be determined. Whether the pair of images have at least one similar object may also be determined, based at least in part on a result of the determination of similarity between the corresponding adjoin features. Other embodiments may be described and claimed.

Abstract: A system for providing positioning functionality in an apparatus. Apparatuses may comprise various sensor resources and may utilize these resources to sense information at a location. For example, an apparatus may sense visual, signal and/or field information at a location. The apparatus may then compare the sensed information to a mapping database in order to determine position.

Abstract: Method and system for obtaining tomosynthesis and material decomposition images of an object of interest using a system comprising an x-ray source facing a detector. The method comprises generating a 2D material decomposition image of an object of interest from at least two sets of acquisitions. Each set is performed at a different energy spectrum and comprises at least one projection image or a plurality of projection images acquired at different x-ray source angulation positions, and the 2D material decomposition image can be generated for a predetermined orientation selected from one of said different x-ray source angulation positions or from a virtual orientation. At least one of the plurality of 3D volume portion images and/or the 2D contrast enhanced material decomposition image are displayed for review by a health care professional.

Abstract: A method of determining interior three-dimensional movement, deformation and strain in an object is provided. The method includes constructing a reference volume image, constructing a deformed volume image, wherein the deformed volume image represents the object after an applied stress and strain. The method further includes comparing the reference volume image and the deformed volume image to determine at least one displacement vector at least one of the x, y, and z directions. In addition it includes making a comparison for size of speckle features measured within the object. Based upon the at least one displacement vector, interior three-dimensional movement, deformation and strain can be measured for the object resulting from the applied stress and strain.

Abstract: A training engine is described which has a memory arranged to access a neural network image classifier, the neural network image classifier having been trained using a plurality of training images from an input space, the training images being labeled for a plurality of classes. The training engine has an adversarial example generator which computes a plurality of adversarial images by, for each adversarial image, searching a region in the input space around one of the training images, the region being one in which the neural network is linear, to find an image which is incorrectly classified into the plurality of classes by the neural network. The training engine has a processor which further trains the neural network image classifier using at least the adversarial images.

Abstract: Systems, computer program products, and techniques for discriminating hand and machine print from each other, and from signatures, are disclosed and include determining a color depth of an image, the color depth corresponding to at least one of grayscale, bi-tonal and color; reducing color depth of non-bi-tonal images to generate a bi-tonal representation of the image; identifying a set of one or more graphical line candidates in either the bi-tonal image or the bi-tonal representation, the graphical line candidates including one or more of true graphical lines and false positives; discriminating any of the true graphical lines from any of the false positives; removing the true graphical lines from the bi-tonal image or the bi-tonal representation without removing the false positives to generate a component map comprising connected components and excluding graphical lines; and identifying one or more of the connected components in the component map.

Abstract: A subtraction image is generated by performing subtraction between a mask image serving as a radiation image obtained by capturing an object, at least of a specific region of which does not include contrast medium, and a live image serving as a radiation image obtained by capturing the object which includes the contrast medium. The emphasis degree serving as the degree of emphasis processing for the subtraction image is determined based on at least either the mask image or live image. The emphasis processing is performed for the subtraction image based on the emphasis degree.

Abstract: A system and method for providing a fast, efficient and accurate method for recovery of depth information from a single image of a face is disclosed. The technique uses a novel thin-plate spline-based dense-correspondence method to align the face, and the representation incorporates a weighted framework, interpreting the depth recovery problem as a weighted data problem.

Abstract: According to an approach for processing image data at network sites, a user invokes functionality on a client device, a scanning device, or a multi-function peripheral (MFP) to cause one or more printed receipts to be scanned to generate image or scan data. The scan data is transmitted by the client device or the MFP to a receipt data management system that maintains access data for a plurality of users and sites. The sites each include a graphical user interface that allows users to manually renter receipt data and request the generation of receipt data processing results. The receipt data management system manages the processing of the receipt data by a specified site on behalf of the user to generate receipt data processing results that satisfy third party requirements. This includes satisfying any authentication requirements of the specified site.

Abstract: The present disclosure relates to image retrieval in an electronic device. An operating method of an electronic device includes determining an object included in an image and a position of extent occupied by the object in the image and retrieving at least one image including the object at the position from among a plurality of images.

Abstract: Data transmission from a bottom hole assembly (BHA) includes obtaining a scan from multiple scans forming a downhole data log of a borehole within a subterranean formation. The scan includes a sequence of data items from a sensor in the BHA located in the borehole. Each data item corresponds to an azimuth angle of the sensor. Further, compressed scan data is generated from the sequence of data items on a per-scan basis, and transmitted, using a pre-determined borehole telemetry, to a surface unit.