Abstract: A method for analyzing images to generate a plurality of output features includes receiving input features of the image and performing Fourier transforms on each input feature. Kernels having coefficients of a plurality of trained features are received and on-the-fly Fourier transforms (OTF-FTs) are performed on the coefficients in the kernels. The output of each Fourier transform and each OTF-FT are multiplied together to generate a plurality of products and each of the products are added to produce one sum for each output feature. Two-dimensional inverse Fourier transforms are performed on each sum.

Abstract: An image processing apparatus which is capable of providing images with enhanced resolution and enhanced sharpness is provided. The image processing apparatus includes an inputter configured to receive an input of image data, an up-scaler configured to up-scale the inputted image data, a region sense unit configured to sense an edge region and a line region within the up-scaled image data by extracting brightness information of the up-scaled image data, an image corrector configured to correct the up-scaled image data in different correction methods for each of the edge region and the line region, and an outputter configured to output the corrected image data.

Abstract: In some implementations, a computer-implemented method is capable of automatically segmenting and detecting a tattoo within an image. An image may be initially obtained. A block coverage pattern that identifies multiple blocks within the obtained image may be determined. A set of processing operations may then be performed for each block. The processing operations may include calculating a plurality of statistical features. A confidence score reflecting a likelihood that at least a portion of the block includes a predetermined graphical attribute associated with tattoos may be calculated. A subset of the multiple blocks of the image that have a respective confidence score greater than a predetermined threshold value may be identified. A portion of the image that includes one or more blocks from among the subset of the multiple blocks may then be determined to correspond to a tattoo.

Abstract: A fingerprint unlocking method and a terminal include acquiring a first fingerprint image. The number of target feature points of the first fingerprint image is determined. A target adjustment parameter corresponding to the number of target feature points is acquired according to a preset mapping relationship between the number of feature points and adjustment parameters. A second fingerprint image is received according to the target adjustment parameter. A terminal is unlocked when the second fingerprint image matches with a preset fingerprint template.

Abstract: Systems, methods, and non-transitory computer-readable media can identify a first video represented based on a first set of image frames. A first subset of image frames can be extracted from the first set of image frames. The first subset of image frames can be compared to one or more image frames associated with a collection of video clusters. It can be determined that at least a threshold quantity of image frames in the first subset matches, within an allowable deviation, at least some image frames associated with a first video cluster included the collection of video clusters. The first video cluster can be defined to include the first video.

Abstract: Embodiments of the invention perform assisted tagging of images, including tagging of people, locations, and activities depicted in those images. A batch of images is received comprising images of faces, including at least some faces that have not yet been tagged. A facial recognition algorithm is applied to the faces to determine matching data comprising possible tags for each untagged face. A logic engine applies logic rules to reduce the likelihood that certain matches are correct. The most likely match from among the possible matches is selected for suggestion to the user for verification. Once verified, the metadata of the image indicating the recognized people within the image is updated.

Abstract: A system for detecting a light emitting apparatus, the system including a light emitting apparatus for emitting a color, a camera configured to obtain a source image that includes a target object having the color, a memory apparatus storing software configured to detect the target object by generating a noise image by adding an additive color to the source image, detecting objects in a YCbCgCr image and an HSV image with respect to the noise image by using respective threshold values of the YCbCgCr image and the HSV image, and combining the YCbCgCr image and the HSV image in which the objects have been detected, and a control circuit configured to detect the target object in the source image by using the software.

Abstract: An image processing device according to an embodiment includes a display control unit, a pattern specification unit, and a pattern collating unit. The display control unit generates a composite image that is a surrounding image of a vehicle viewed from a virtual viewpoint, based on images input from in-vehicle cameras, and displays thereof on a display unit. The pattern specification unit specifies an image pattern to be output as the input image among image patterns, for each in-vehicle camera. The pattern collating unit determines whether all image pattern responses from the in-vehicle cameras match with the image pattern. The display control unit displays a certain preparation image on the display unit when the display switching condition to the composite image is satisfied, and displays the composite image thereon instead of the preparation image, when the pattern collating unit determines that all the image pattern responses match with the image pattern.

Abstract: An imaging assembly includes a radiation camera configured to obtain position-sensitive radiation data of at least one source of interest. The assembly includes a controller includes a processor and tangible, non-transitory memory on which is recorded instructions for executing a method for reducing background interference in the position-sensitive radiation data in an image domain. The controller is programmed to generate a first set of image data (G1) from the position-sensitive radiation data with a first energy window. A second set of image data (G2) is generated from the position-sensitive radiation data with a second energy window such that the second energy window does not overlap with the first energy window. A third set of image data (G3) is generated having reduced background interference, based at least partially on the first and second sets of image data.

Abstract: Embodiments of the present invention provide a system for estimating a location of a gas leak, based on machine learning from forward gas concentration data provided by an analog or scale model including a gas source. The system improves significantly over previous systems by providing high quality, physically accurate forward modeling data inexpensively. During operation, the system configures an aerosol source at a first location to emit a gaseous aerosol. The system then configures a laser source to illuminate the aerosol with a laser sheet. The system may then obtain an image of a reflection of the laser sheet from the aerosol. The system may then analyze the image to quantify a three-dimensional concentration distribution of the aerosol. The system may then estimate, based on solving an inverse problem and an observed second gas concentration, a second location of a second gas source.

Abstract: An apparatus and a method for searching for an image are provided. The apparatus includes an image receiver configured to receive an image to be searched for; an image processor configured to detect an edge of the received image and to filter pixels included on the edge of the received image based on parameters related to gradients of the edge of the received image; and an image search unit configured to search for one or more images related to the image to be searched for based on the filtered pixels included on the edge of the received image.

Abstract: A fingerprint recognition apparatus includes an electrode-and-wiring substrate having two main surfaces opposite to each other, where one main surface is in proximity to user finger and the electrode-and-wiring substrate has a plurality of sensing electrodes on the other main surface. The fingerprint recognition apparatus further includes an integrated circuit (IC) chip having a fingerprint sensing circuit and a plurality of metal bumps. At least part of the metal bumps are electrically connected to the fingerprint sensing circuit and corresponding sensing electrodes on the electrode-and-wiring substrate, whereby the fingerprint sensing circuit is electrically connected to the sensing electrodes.

Abstract: A method of establishing a 3D saliency model based on 3D contrast and depth weight, includes dividing left view of 3D image pair into multiple regions by super-pixel segmentation method, synthesizing a set of features with color and disparity information to describe each region, and using color compactness as weight of disparity in region feature component, calculating feature contrast of a region to surrounding regions; obtaining background prior on depth of disparity map, and improving depth saliency through combining the background prior and the color compactness; taking Gaussian distance between the depth saliency and regions as weight of feature contrast, obtaining initial 3D saliency by adding the weight of the feature contrast; enhancing the initial 3D saliency by 2D saliency and central bias weight.

Abstract: A Convolutional Neural Network (CNN) includes an initial set of convolutional layers and max pooling units, in which any input is convoluted with the learned image filters and the output is a stack of the different filter responses. Max pooling produces a scaled version of the output. The process can be repeated several times, resulting in a stack of space invariant-scaled images. Since the operation is space invariant, the computations of these layers not need to be recomputed if interested just in certain regions of the image. A Region Of Interest (ROI) Pooling layer is used to select regions to be processed by the set of fully connected layers, which uses the response of the multiple convolutional layers of the network to determine the regions where the objects (of different scales) could be located. This object proposal method is implemented as a Region Of Interest (ROI) Selector.

Abstract: An optical module includes a housing with a lighting mechanism, an aperture formed therein, and a window that frames a transparent surface adapted to contact a skin surface of a person, for example, a fingertip. The lighting mechanism provides light to illuminate the skin surface placed upon the transparent surface. A prism has a first side facing the lighting mechanism, a second side at the window, and a third side through which a dactylographic image exits the prism. One or more light reflecting surfaces are disposed within the housing to reflect the dactylographic image towards the housing aperture. The optical module is coupled to a mobile device having a camera, with the aperture of the housing aligning with a lens of the camera. The camera acquires the dactylographic image, and the mobile device adjusts this dactylographic image to produce a dactylographic image suitable for biometric matching.

Abstract: A sensor screen and a display device including the same are discussed. The sensor screen can include a first base member including a first active area and a first bezel area and having a first thickness, a second base member positioned opposite the first base member and having a second thickness less than the first thickness; and a fingerprint sensor disposed on one of the first and second base members which are positioned opposite each other. The fingerprint sensor can include a first group of fingerprint electrodes, which are arranged in a first direction in a portion of the first active area, and a second group of fingerprint electrodes, which are insulated from the first group of fingerprint electrodes and are arranged in a second direction crossing the first direction.

Abstract: There is provided a method of storing a set of biometric data templates, which emanates from a biometric object, among a plurality of biometric data templates, wherein each biometric data template comprises biometric data from only a part of a biometric object. The method comprises aligning biometric data templates with each other in view of the biometric object of the plurality of biometric data templates which comprises partially overlapping biometric data with at least one other biometric data template; determining the set of biometric data templates to be stored; and storing the set of biometric data templates.

Abstract: Video and corresponding metadata is accessed. Events of interest within the video are identified based on the corresponding metadata, and best scenes are identified based on the identified events of interest. Events of interest can be tagged within the video based on, for instance, user input, audio signals, motion vectors, and metadata corresponding to the video. A camera system can process video data based on the events of interest tagged within the video before outputting the video data. For instance, video scenes associated with tagged events of interest can be combined to form a video highlight clip. Likewise, portions of video tagged with events of interest can be encoded or stored at a higher resolution or frame rate than other portions of the video.

Abstract: A computer-implemented method executes instructions for receiving or retrieving image files. A first image file corresponds to both a scaling article image having one or more known dimension(s) and to a target article image submitted by a consumer or a vendor. Contours are determined for each article image to obtain one or more pixel dimension(s) for each article image. The pixel dimension(s) for the scaling article image are related to its known dimension(s) to obtain scaled pixel dimensions for the target article image. A consumer profile is generated which includes at least one of a set of curves that represent the determined contours and a set of control points, measure lines and control lines to represent the measurement of the reference article. Access to the profile facilitates creation of new articles for the consumer and/or assessment of how an article having a similarly derived profile will fit.

Abstract: According to the present disclosure, an image-processing apparatus acquires horizontal black hat image data by taking a difference between source image data and horizontal closing image data obtained by performing closing processing on each pixel row of the source image data, acquires vertical black hat image data by taking a difference between the source image data and vertical closing image data obtained by performing the closing processing on each pixel column of the source image data, and acquires logical conjunction image data where a logical conjunction of the horizontal black hat image data and the vertical black hat image data is taken.