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 system is provided for adaptively conformed imaging of work pieces having disparate configurations, which comprises a mount unit for holding at least one work piece, and an imaging unit for capturing images of the work piece held by the mount unit. At least one manipulator unit is coupled to selectively manipulate at least one of the mount and imaging units for relative displacement therebetween. A controller coupled to automatically actuate the manipulator and imaging units executes scene segmentation about the held work piece, which spatially defines at least one zone of operation in peripherally conformed manner about the work piece. The controller also executes an acquisition path mapping for the work piece, wherein a sequence of spatially offset acquisition points are mapped within the zone of operation, with each acquisition point defining a vantage point for the imaging unit to capture an image of the work piece from.

Abstract: A method and apparatus for (a) receiving a three-dimensional (3D) point cloud from a depth sensor that is oriented towards an open end of a shipping container, the point cloud comprising a plurality of points that each have a respective depth value, (b) segmenting the received 3D point cloud among a plurality of grid elements, (c) calculating a respective loaded-container-portion grid-element volume for each grid element, (d) calculating a loaded-container-portion volume of the shipping container by aggregating the calculated respective loaded-container-portion grid-element volumes, (e) calculating an estimated fullness of the shipping container based on the loaded-container-portion volume and a capacity of the shipping container; and (f) outputting the calculated estimated fullness of the shipping container.

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: Object identification through image matching can utilize ratio and other data to accurately identify objects having relatively few feature points otherwise useful for identifying objects. An initial image analysis attempts to locate a “scalar” in the image, such as may include a label, text, icon, or other identifier that can help to narrow a classification of the search, as well as to provide a frame of reference for relative measurements obtained from the image. By comparing the ratios of dimensions of the scalar with other dimensions of the object, it is possible to discriminate between objects containing that scalar in a way that is relatively robust to changes in viewpoint. A ratio signature can be generated for an object for use in matching, while in other embodiments a classification can identify priority ratios that can be used to more accurately identify objects in that classification.

Abstract: A service can perform optical character recognition (OCR) on an image of a record to determine a first set of information items about the record. A second set of information items can be identified that are likely part of the record but not determinable from performing OCR on the image. Another resource can be utilized to determine the second set of information items. A classification for the record can be determined based on first and second sets of information items. The record can be associated with a financial resource of the user based at least in part on the classification.

Abstract: A display control method executed by a processor included in a terminal device that includes a display and a camera disposed at a display surface side of the display, the display control method includes acquiring a first image taken by the camera, the first image including a second image of a face; calculating a proportion of a region of the second image to a region of the first image; determining whether the proportion is equal to or higher than a threshold; displaying the first image on the display when it is determined that the proportion is equal to or higher than the threshold; and restraining the first image from being displayed on the display when it is determined that the proportion is not equal to or higher than the threshold.

Abstract: Various embodiments described herein notifying users regarding photos in which they may appear and suggest photo tags accordingly. Subject to user preferences and privacy settings, facial recognition with respect to a specific user of a social networking system may be performed on one or more photos added by, or otherwise associated with, other entities of the social networking system. For those photos in which the specific user is facially recognized, a suggested photo tag for the specific user may be associated with the recognized photos and the specific user may be alerted accordingly. Depending on the embodiment, the specific user may be provided with an option to confirm the suggested photo tag, decline the suggested photo tag, or do nothing. In the event the specific user declines the suggested photo tag with respect to a particular photo, other users may be prevented from tagging the specific user with respect to the particular photo.

Abstract: Method of authentication of a user via a terminal including acquisition elements for performing a biometric acquisition, includes: presentation to the acquisition elements by the user of a member to be verified by biometric comparison and triggering by the user a biometric acquisition to obtain an acquisition datum; biometric comparison of the acquisition datum with a prestored reference datum, the user being authenticated if the biometric comparison is positive and not authenticated otherwise, the method further including the following steps: providing the user with elements indicative of a first angle, the user applying a corrective rotation by a second angle so as to match an acquisition datum with the second angle, and in the biometric comparison step the reference datum is matched with the first angle, the biometric comparison being positive if the second angle is substantially equal to the first angle. Device adapted to implement such a method.

Abstract: Systems, methods, devices, media, and computer readable instructions are described for local image tagging in a resource constrained environment. One embodiment involves processing image data using a deep convolutional neural network (DCNN) comprising at least a first subgraph and a second subgraph, the first subgraph comprising at least a first layer and a second layer, processing, the image data using at least the first layer of the first subgraph to generate first intermediate output data; processing, by the mobile device, the first intermediate output data using at least the second layer of the first subgraph to generate first subgraph output data, and in response to a determination that each layer reliant on the first intermediate data have completed processing, deleting the first intermediate data from the mobile device. Additional embodiments involve convolving entire pixel resolutions of the image data against kernels in different layers if the DCNN.

Abstract: Methods and apparatus are provided for generating computer 3D models of an object, by registering two or more scans of physical models of an object. The scans may be 3D scans registered by a curve-based registration process. A method is provided for generating a 3D model of a portion of a patient's oral anatomy for use in dental restoration design. Also provided are scanning workflows for scanning physical models of an object to obtain a 3D model.

Abstract: A touch display device and a touch method thereof are provided. The touch display device includes a touch panel, a fingerprint collecting unit, a signal processing unit and a memory. The fingerprint collecting unit is connected with the touch panel, and the signal processing unit is connected with the touch panel and the memory. The touch method of the touch display device includes acquiring a fingerprint information set including M first fingerprint images by the fingerprint collecting unit; and inquiring M second fingerprint images matched with the M first fingerprint images in a preset information set according to the M first fingerprint images in the fingerprint information set in the memory by the signal processing unit, and acquiring S first control instructions corresponding to the M second fingerprint images from the memory.

Abstract: Provided is an apparatus for measuring a quality of a holographic image. The apparatus includes: a hologram image reconstruction unit configured to perform a hologram image; a superimposed focus image generation unit configured to collect an area that is in focus from reconstructed images and superimpose into a single image; and a hologram image quality measurement unit configured to perform an image quality evaluation for the superimposed focus image.

Abstract: A fingerprint sensing system has a fingerprint sensing device and a power supply circuit. The fingerprint sensing device has a high-voltage input terminal and a low-voltage input terminal. During scan phases of the fingerprint sensing device, the power supply circuit provides a first voltage to the high voltage input terminal and provides a second voltage to the low voltage input terminal. During the read phases of the fingerprint sensing device, the power supply circuit provides a third voltage to the high voltage input terminal and provides a fourth voltage to the low voltage input terminal. A first voltage difference is between the first voltage and the second voltage. A second voltage difference is between the third voltage and the fourth voltage. The first voltage difference is greater than the second voltage difference.

Abstract: An image recognition (IR) computing device is provided herein, the IR computing device configured to receive a search request from a user computing device instructing the IR computing device to conduct a search for a subject clothing item pictured in a subject image, the search request including the subject image. The IR computing device is further configured to analyze the subject image and compare a plurality of vendor images to the subject image. The IR computing device may employ an object recognition component to analyze and compare the images. The IR computing device is further configured to generate a list of potential matches to the subject image, transmit the list of potential matches to the user computing device for display to a user within a clothing match app, and receive an indication from the user computing device of whether the search was successful or unsuccessful.

Abstract: An image processing apparatus includes a first obtaining unit to obtain a number of images in which a pair of a first person and a second person simultaneously appears, out of a plurality of images. A detection unit detects whether the pair appears in a target image group, which is one of a plurality of image groups into which the plurality of images is classified. In determining whether the pair appears in the target image group, it is determined that the pair appears in the target image group not only in a case when the target image group includes an image in which the first person and the second person simultaneously appear, but also, in a case when the target image group includes images in which the first person and the second person appear separately.

Abstract: Methods and systems for recognizing people in images with increased accuracy are disclosed. In particular, the methods and systems divide images into a plurality of clusters based on common characteristics of the images. The methods and systems also determine an image cluster to which an image with an unknown person instance most corresponds. One or more embodiments determine a probability that the unknown person instance is each known person instance in the image cluster using a trained cluster classifier of the image cluster. Optionally, the methods and systems determine context weights for each combination of an unknown person instance and each known person instance using a conditional random field algorithm based on a plurality of context cues associated with the unknown person instance and the known person instances. The methods and systems calculate a contextual probability based on the cluster-based probabilities and context weights to identify the unknown person instance.

Abstract: A method of matching images A and B of the same scene taken at different locations in the scene is provided by matching correspondence points in the image by evaluating pixel characteristics from nearby regions using a constellation of image chips and utilizing joint information across multiple resolution levels in a probability framework. Since each image chip is small, each chip in one image potentially can be matched with a number of chips in the other image. The accumulation of evidence (probability) over all image chips within the constellation over multiple resolution levels reduces the ambiguity. The use of a constellation of image chips removes the requirement present in most visual point matching techniques to special feature points (e.g. corner points) as the correspondence points.

Abstract: Methods and systems for finding patterns in a design for a specimen are provided. One system includes one or more computer subsystems configured for searching for a target pattern in a design for a specimen to thereby find multiple instances of the target pattern in the design. The one or more computer subsystems are also configured for separating the multiple instances of the target pattern into different groups based on information for surrounding patterns within a predefined window around the target pattern such that each of the different groups corresponds to a different combination of the target pattern and the surrounding patterns.