Abstract: A set of data comprising at least a first tile and a second tile is received. A determination of a measure is made, for a first pixel location included in both the first tile and the second tile, of whether the first pixel location corresponds to a boundary. A boundary map is generated as output. The boundary map can be used, in conjunction with an image, to regularize pixel brightness values of the image.

Abstract: An artificial neural network for analyzing input data, the input data being a 3D tensor having D channels, such as D frames of a video snippet, to recognize an action therein, including: D spatial transformer modules, each generating first and second spatial transformations and corresponding first and second attention windows using only one of the D channels, and transforming first and second regions of each of the D channels corresponding to the first and second attention windows to generate first and second patch sequences; first and second CNNs, respectively processing a concatenation of the D first patch sequences and a concatenation of the D second patch sequences; and a classification network receiving a concatenation of the outputs of the first and second CNNs and the D sets of transformation parameters of the first transformation outputted by the D spatial transformer modules, to generate a predicted action class.

Abstract: An apparatus for generating precision maps of an area is disclosed. The apparatus receives sensor data, where the sensor data includes sensor readings each indicating a level of a parameter in one of a plurality of first portions of an area, and video data representing an aerial view of the area. The sensor data may be received from sensors that are each deployed in one of the first portions of the area. The video data may be received from an aerial vehicle. An orthomosaic may be generated from the video data, and the orthomosaic and the sensor data used to generate a predication model. The prediction model may then be used to extrapolate the sensor data to determine a level of the parameter in each of a plurality of second portions of the area. A precision map of the area may be generated using the extrapolated sensor readings.

Abstract: The invention has for object a device for controlling (2) at least one audio or video signal comprising: means for controlling mounted mobile in rotation about an axis of rotation (z) on a base, first means of detecting a displacement in rotation of the means for controlling able to generate a first signal, said first signal supplying means for processing said at least one audio or video signal, with the first means of detecting comprising means for measuring the angle of rotation of the means for controlling, means for displaying and/or light-emitting means, comprising several sources of light forming at least first and second means of graduation. According to the invention, said sources of light of said first and second means of graduation are selectively controlled by said means for processing according to the measurement of the angle of the rotation of said means for controlling.

Abstract: An image data processor (106) includes a structural image data processor (114) that employs a multi-structure atlas to segment a region of interest from structural image data that includes tissue of interest and that segments the tissue of interests from the region of interest. The image data processor further includes functional image data processor (116) that identifies the tissue of interest in functional image data based on the segmented tissue of interest. An image data processor includes a multi-structure atlas generator (104) that generates a multi-structure atlas. The multi-structure atlas physically maps structure to tissue of interest such that locate the structure in structural image data based on the multi-structure atlas localizes the tissue of interest to the region of interest.

Abstract: The present application involves a railroad track asset surveying system comprising an image capture sensor, a location determining system, and an image processor. The image capture sensor is mounted to a railroad vehicle. The location determining system holds images captured by the image capture sensor. The image processor includes an asset classifier and an asset status analyzer. The asset classifier detects an asset in one or more captured images and classifies the detected asset by assigning an asset type to the detected asset from a predetermined list of asset types according to one or more features in the captured image. The asset status analyzer identifies an asset status characteristic and compares the identified status characteristic to a predetermined asset characteristic so as to evaluate a deviation therefrom.

Abstract: An image processing apparatus includes an acquisition unit, a first calculation unit, a combination unit, a second calculation unit, and a correction unit. The acquisition unit acquires a plurality of pieces of image data captured sequentially in time series. The first calculation unit calculates a first displacement between one piece of image data among the plurality of pieces of image data and a piece of image data adjacent in time series to the one piece of image data. The combination unit combines a plurality of first displacements to obtain a second displacement between the one piece of image data and a piece of image data neighboring the one piece of image data in time series. The second calculation unit calculates a correction amount for the one piece of image data based on the second displacement. The one piece of image data is corrected based on the calculated correction amount.

Abstract: Innovations in the identification or classification of features in a data set are described, such as a data set representing measurements taken by a scientific instrument. For example, a task-specific processing component, such as a video encoder, is used to generate task-specific metadata. When the data set includes video frames, metadata can include information regarding motion of image elements between frames, or other differences between frames. A feature of the data set, such as an event, can be identified or classified based on the metadata. For example, an event can be identified when metadata for one or more elements of the data set exceed one or more threshold values. When the feature is identified or classified, an output, such as a display or notification, can be generated.

Abstract: Aspects disclosed herein provide methods, systems, and devices for 3D image capture based on a Structure from Motion (SFM) imaging technique with simplified pose detection. The present disclosure provides a straightforward method to directly track a camera's motion (pose detection) thereby removing a substantial portion of the computing load needed to build a dimensionally accurate 3D model from a sequence of images taken by the camera.

Abstract: The invention notably relates to a computer-implemented method for localizing a 3D modeled object in a 3D scene, the method comprising a positioning of the 3D modeled object in the 3D scene, the positioning being performed following an algorithm that rewards, for each of first couples made of two 3D points of the 3D modeled object and their respective associated normal vectors, a match with a respective second couple made of two 3D point of the 3D scene and its respective associated normal vectors, the match between the first couple and the second couple amounting to a substantial equality between the value of a descriptor for the first couple and the value of the descriptor for the second couple, the descriptor being variable under mirror symmetries. This improves the localizing a 3D modeled object in a 3D scene.

Abstract: An image processing method to obtain a high sense of depth or high stereoscopic effect for an image and a display device utilizing the method are provided. Image data of an image is separated into image data of a plurality of objects and a background. A feature amount is obtained from the image data of each object, so that the objects are identified. The relative distance between viewer's eye and any of the objects is determined by the data of the sizes of the objects in the image and the sizes of the objects stored in the database. The image data of each object is processed so that an object with a shorter relative distance is enlarged. The image data of each object after image processing is combined with the image data of the background, so that a sense of depth or stereoscopic effect of an image is increased.

Abstract: Accelerating object detection techniques are described. In one or more implementations, adaptive sampling techniques are used to extract features from an image. Coarse features are extracted from the image and used to generate an object probability map. Then, dense features are extracted from high-probability object regions of the image identified in the object probability map to enable detection of an object in the image. In one or more implementations, cascade object detection techniques are used to detect an object in an image. In a first stage, exemplars in a first subset of exemplars are applied to features extracted from the multiple regions of the image to detect object candidate regions. Then, in one or more validation stages, the object candidate regions are validated by applying exemplars from the first subset of exemplars and one or more additional subsets of exemplars.

Abstract: A method for generating an improved stack from wide azimuth data. This is accomplished by displaying gather traces in a geometrically sensible way that depends on gather and processing type. Once a gather is available in such a manner, a geoscientist can use various geometric or freehand shapes to indicate which traces from the gather to stack and which to ignore. Given sufficient computing power, this can be done interactively, allowing a geoscientist to determine a set of traces that create an optimal stack at a single location in the output cube. Multiple locations can be improved in this manner and a final stack can be generated that contains a blended agglomeration of all of the improvements.

Abstract: Provided are a fingerprint sensor and a display device including the same. A fingerprint sensor array includes: a plurality of first electrodes arranged in a first direction, each of the plurality of first electrodes including at least one of: a first transparent electrode including a transparent conductive material and a first metal electrode including a low-resistance metal material; a plurality of second electrodes arranged in a second direction crossing the first direction, each of the plurality of second electrodes including at least one of: a second transparent electrode including the transparent conductive material and a second metal electrode including the low-resistance metal material; and an insulating layer between the plurality of first electrodes and the plurality of second electrodes to insulate the first electrodes from the second electrodes.

Abstract: Disclosed herein is a method of optimizing data normalization by selecting the best height normalization setting from training RNN (Recurrent Neural Network) with one or more datasets comprising multiple sample images of handwriting data, which comprises estimating a few top place ratios for normalization by minimizing a cost function for any given sample image in the training dataset, and further, determining the best ratio from the top place ratios by validating the recognition results of sample images with each top place ratio.

Abstract: An improved eye tracker system and methods for detecting eye parameters including eye movement using a pupil center, pupil diameter (i.e., dilation), blink duration, and blink frequency, which may be used to determine a variety of physiological and psychological conditions. The eye tracker system and methods operates at a ten-fold reduction in power usage as compared to current system and methods. Furthermore, eye tracker system and methods allows for a more optimal use in variable light situations such as in the outdoors and does not require active calibration by the user.

Abstract: A learning method for acquiring a bounding box corresponding to an object in a training image from multi-scaled feature maps by using a CNN is provided. The learning method includes steps of: (a) allowing an N-way RPN to acquire at least two specific feature maps and allowing the N-way RPN to apply certain operations to the at least two specific feature maps; (b) allowing an N-way pooling layer to generate multiple pooled feature maps by applying pooling operations to respective areas on the at least two specific feature maps; and (c) (i) allowing a FC layer to acquire information on pixel data of the bounding box, and (ii) allowing a loss layer to acquire first comparative data, thereby adjusting at least one of parameters of the CNN by using the first comparative data during a backpropagation process.

Abstract: As the use of facial biometrics expands in the commercial and government sectors, the need to ensure that human facial examiners use proper procedures to compare facial imagery will grow. Human examiners have examined fingerprint images for many years such that fingerprint examination processes and techniques have reached a point of general acceptance for both commercial and governmental use. The growing deployment and acceptance of facial recognition can be enhanced and solidified if new methods can be used to assist in ensuring and recording that proper examination processes were performed during the human examination of facial imagery.

Abstract: Various aspects of an image-processing system and method to reconstruct a three-dimensional (3D) anatomical surface of an anatomical portion are disclosed herein. The system includes an image-processing device configured to receive a plurality of stereo images of the anatomical portion. A first set of key points with a point density greater than a threshold value is identified. A second set of key points is determined based on filtration of one or more outlier key points from the identified first set of key points. A 3D anatomical surface of the anatomical portion is reconstructed based on disparity of one or more matched key points in the determined second set of key points and a smoothing operation performed on the disparity.

Abstract: A system and method of estimating the body shape of an individual from input data such as images or range maps. The body may appear in one or more poses captured at different times and a consistent body shape is computed for all poses. The body may appear in minimal tight-fitting clothing or in normal clothing wherein the described method produces an estimate of the body shape under the clothing. Clothed or bare regions of the body are detected via image classification and the fitting method is adapted to treat each region differently. Body shapes are represented parametrically and are matched to other bodies based on shape similarity and other features. Standard measurements are extracted using parametric or non-parametric functions of body shape. The system components support many applications in body scanning, advertising, social networking, collaborative filtering and Internet clothing shopping.