Abstract: Computerized methods and systems, including computer programs encoded on a computer storage medium, may detect events shown within digital video content captured by one or more video cameras during a prior time period, and predict an occurrence of an additional event during a future time period based on time-varying patterns among the detected events. For example, a computing system may detect events shown within digital video content captured by one or more video cameras, and may establish a predictive model that identifies one or more time-varying patterns in event parameter values that characterize the detected events within the prior time period. Based on an outcome of the predictive model, the computing system may determine an expected value of one of the event parameters during the second time period.

Abstract: A sensitive measure of brain condition simultaneously evaluates multiple measurements of water diffusion in brain tissue combined so as to correct for covariance between the different data types of the multipoint measurements and compares the multipoint measurements to a corresponding multipoint measure representing normal brain tissue to provide a distance indicating a likelihood of atypical brain conditions.

Abstract: The present invention relates to a method of quantifying a lean tissue volume comprising the steps of acquiring (10) a acquired image as a water-fat separated magnetic resonance image, wherein the acquired image comprises a water image and a fat image, providing (20) a calibrated fat image (F), providing (30) a soft tissue mask (STM) defining areas of soft tissue in the acquired image, and defining (40) a region of interest (ROI) of the acquired image. The method further comprises a step of calculating (50) a lean tissue volume (LTV) by multiplying, for each volume element in the region of interest, the soft tissue mask with the volume (Vvox) of each volume element and the result of one minus the calibrated fat image, and summarizing the products of said multiplications for all volume elements in the region of interest.

Abstract: The method of generating multi-parametric nosological images including the steps of obtaining a stack of images; improving such images; reducing the images to a new space of equivalent representation; classifying in an unsupervised manner the units of the images obtained according to a pre-established number of classes without biological interpretation; automatically assigning multi-parametric profiles to the obtained classes; and generating a nosological image through the obtained classes. The system includes an imaging unit; an image processing unit for processing the obtained images and generating therefrom a multi-parametric nosological image; a connection for transmitting the images obtained by the imaging unit to the image processing unit; and a display for displaying the generated multi-parametric nosological image.

Abstract: In an embodiment of the invention, a plurality of images of a region under examination are recorded at different times. Anatomical information and flow information are derived from the images. The anatomical information may relate to the course of vessels or the structure of perfused tissue. The temporal component of the flow information can advantageously be combined with the anatomical information in a resultant image. An intensity-dependent fenestration assigns a gray-scale value to pixels of the resultant image in accordance with the anatomical information. A time-dependent fenestration assigns a chromaticity to the pixels of the resultant image in accordance with the flow information and the gray-scale values and the chromaticities are assigned independently of one another.

Abstract: Method and system is disclosed for image segmentation. The method includes acquiring a digital image, constructing a directed graph from the digital image, calculating a plurality of cost functions, constructing an electrical network based upon the constructed directed graph and the plurality of calculated cost functions, simulating the electrical network using fixed-point linearization, and thresholding the voltages in the simulated electrical network to produce image segmentation. Simulation may be executed in parallel to achieve desirable computational efficiencies.

Abstract: A patient image record and a reference image record are read for capturing anatomical features of a patient on the basis of medical image records is provided. Joints are localized in the patient image record and in the reference image record by means of a pattern recognition method. At the localized joints, the patient image record is segmented into patient image segments and the reference image record is segmented into reference image segments. According to the invention, a segment-individual image deformation is established for a respective reference image segment or patient image segment, by means of which image deformation image structures of the reference image segment are mapped to corresponding image structures of the patient image segment). Then, an anatomical feature is captured on the basis of the segment-individual maps between the reference image segments) and the patient image segments.

Abstract: Information indicating the correspondence relationship of imaging conditions, the thickness of an object, and the thickness of a specific composition included in the object, and a contrast correction amount is stored. The thickness of the object and the thickness of a specific composition included in the object of each unit region having one or two or more pixels of the radiation image are acquired. For each unit region of the radiation image, a contrast correction amount corresponding to a combination of the imaging conditions, the thickness of the object acquired for the unit region, and the thickness of the specific composition is acquired with reference to the stored information indicating the correspondence relationship, and contrast correction is performed using the acquired contrast correction amount.

Abstract: An apparatus for food search service includes a food region extractor configured to perform detection in regions in an image where food is present and extract a plurality of candidate regions; a candidate region refiner configured to cluster the candidate regions into groups according to a ratio of overlap between the candidate regions; and a search result generator configured to determine a position of a food region and a food item from the grouped candidate regions.

Abstract: A computer-implemented method for detecting objects by using subcategory-aware convolutional neural networks (CNNs) is presented. The method includes generating object region proposals from an image by a region proposal network (RPN) which utilizes subcategory information, and classifying and refining the object region proposals by an object detection network (ODN) that simultaneously performs object category classification, subcategory classification, and bounding box regression. The image is an image pyramid used as input to the RPN and the ODN. The RPN and the ODN each include a feature extrapolating layer to detect object categories with scale variations among the objects.

Abstract: An image recognition approach employs both computer generated and manual image reviews to generate image tags characterizing an image. The computer generated and manual image reviews can be performed sequentially or in parallel. The generated image tags may be provided to a requester in real-time, be used to select an advertisement, and/or be used as the basis of an internet search. In some embodiments generated image tags are used as a basis for an upgraded image review. A confidence of a computer generated image review may be used to determine whether or not to perform a manual image review.

Abstract: A system and method is provided for identifying an object along a road, where the object may be represented by a bounding box, and projecting a set of obstacle points within the bounding box corresponding to the identified object. In one aspect, a two-dimensional plane oriented perpendicular to a direction of the movement of the vehicle may be identified. In another aspect, the areas of the plane that may be occupied based on the set of obstacle points may be determined to generate a contour of the identified object. Thereafter, the height profiles of the identified object and the vehicle may be determined and identified, respectively. Based on the height profiles, a minimum clearance may be determined.

Abstract: A system for monitoring a patient in a patient area having one or more detection zones, the system comprising one or more cameras, a user interface, and a computing system configured to receive a chronological series of frames from the one or more cameras, identify liquid candidates by comparing a current frame with a plurality of previous frames of the chronological series, determine locations of the liquid candidates, identify thermal signatures of the liquid candidates, determine types of liquids of the liquid candidates based on the locations and thermal signatures of the liquid candidates, and generate an alert with the user interface corresponding to the determined types of liquids.

Abstract: A vision system for a vehicle includes a forward viewing camera. Responsive at least in part to processing by an image processor of image data captured by the forward viewing camera, at least one road characteristic of a road along which the vehicle is traveling is detected. Responsive at least in part to processing by the image processor of image data captured by the forward viewing camera, other vehicles exterior the equipped vehicle are detected. Responsive at least in part to processing by the image processor of image data captured by the forward viewing camera, road curvature of the road along which the vehicle is traveling is determined. Data derived from image data captured by the forward viewing camera may be wirelessly communicated from the vehicle to a remote data receiver located remote from the vehicle.

Abstract: To make a user easily obtain an objective and stable analysis result of bone mineral density. An analyzer 100 of bone mineral density using CT image data of a phantom having a known bone mineral density includes: a known data storage part 105 that stores known data of bone mineral density for a phantom; a histogram production part 102 that produces a histogram of region number relative to a CT value for three-dimensional CT image data of the phantom; a correspondence determination part 106 that determines correspondence between a CT value and a bone mineral density by correlating CT values showing respective peaks of the produced histogram with the known data of the phantom; and an analysis part 109 that decides a bone mineral density for three-dimensional CT image data of a subject using the determined correspondence.

Abstract: A computerized method for determining an objective tumor response to an anti-cancer therapy using cross-sectional images includes receiving a cross-sectional image of digital medical image data, identifying a set of pixels within the cross-sectional image corresponding to a target lesion, and analyzing the set of pixels with an image processing module to identify an upper and lower threshold of pixel intensities within a total range of pixel intensities, define a subset of pixels from the set of pixels that correspond to vascularized tumor and that have pixel intensities greater than or equal to the lower threshold and lower than or equal to the upper threshold, and derive a vascular tumor burden for the set of pixels. The method further includes determining the objective tumor response for the target lesion, wherein the objective tumor response is based on the vascular tumor burden.

Abstract: Digital watermarking is adapted for the variable data printing. A reference signal serves as a proxy for optimizing the embedding a watermark in a host image to be printed. Using the reference signal, embedding parameters are generated, which are a function of constraints such as visual quality and robustness of the machine readable data. Adjustments needed to embed a unique payload in each printed piece are generated using the embedding parameters. These adjustments are stored in a manner that enables them to be efficiently obtained and applied within the RIP or press during operation of the press. Various other methods, system configurations and applications are also detailed.

Abstract: A method and device for analyzing and treating tonal imperfections on human skin. The method has the steps of taking at least one background image of at least 10 ?m2 of skin and calculating the average background L value of the image on a grey scale. One or more treatment images of the skin are taken and at least one localized L value of individual pixels or group of pixels is calculated within the treatment image. The local L value is compared to the background L value. Skin deviations are areas of skin where the absolute value of the difference between a local L and the background L is greater than a predetermined ?LS. Skin deviations are identified by this method and then treated with a treatment composition.

Abstract: A method and device for analyzing and treating tonal imperfections on human skin. The method has the steps of providing a preselected background L value. One or more treatment images of the skin are taken and at least one localized L value of individual pixels or group of pixels is calculated within the treatment image. The local L value is compared to the background L value. Skin deviations are areas of skin where the absolute value of the difference between a local L and the background L is greater than a predetermined ?LS. Skin deviations are identified by this method and then treated with a treatment composition.

Abstract: Video is often used to count and identify people entering and exiting an area, such as a store. This information may be used to determine the number and amount of time people are in the area, and other statistics and characteristics. These variables may be skewed by the associates employed or working in the area, instead of patrons or others. The present disclosure is directed to method for eliminating associates from video by facial analysis and the application of analytics.