Abstract: A method is provided for classifying an image. The method includes inferring location information of an object of interest in an input representation of the image. The method further includes determining foreground object features and background object features from the input representation of the image. The method additionally includes pooling the foreground object features separately from the background object features using the location information to form a new representation of the image. The new representation is different than the input representation of the image. The method also includes classifying the image based on the new representation of the image.
Type:
Grant
Filed:
November 14, 2012
Date of Patent:
June 24, 2014
Assignee:
NEC Laboratories America, Inc.
Inventors:
Olga Russakovsky, Yuanqing Lin, Kai Yu, Fei-Fei Li
Abstract: The invention relates to video surveillance systems that are used, for example, for surveying public places, stations, streets, industrial estates, buildings or similar. Said video surveillance systems comprise one or more surveillance cameras that are oriented towards surveillance scenes and transfer image data streams, in the form of image sequences, to an evaluation center. The invention also relates to a device (1) for identifying and/or classifying a movement pattern in an image sequence of a surveillance scene comprising a plurality of moving objects, an interface (2) for recording the image sequence, a calculation module (5) for determining an optical flow field (10) in the surveillance scene by evaluating the image sequence and an identification module (6) that is designed in terms of programming and/or circuitry such that the optical field and/or partial areas thereof are compared to one or more patterns in order to identify the movement pattern in the image sequence.
Type:
Grant
Filed:
June 6, 2008
Date of Patent:
June 24, 2014
Assignee:
Robert Bosch GmbH
Inventors:
Julia Ebling, Hartmut Loos, Matthias Koenig
Abstract: A system, circuit, and methods for increasing speed of an asynchronous pulse processing based hyperspectral target detection algorithm are disclosed. Stored optimized filter coefficients are used to provide initial filter coefficients. The initial filter coefficients are optimized using an asynchronous pulse processor based hyperspectral detection algorithm to provide optimized filter coefficients, and the optimized filter coefficients are stored.
Type:
Grant
Filed:
October 24, 2011
Date of Patent:
June 17, 2014
Assignee:
The Boeing Company
Inventors:
Vikas Kukshya, Roy Matic, Peter Petre, David R. Gerwe
Abstract: The invention may include a fingerprint scanner having an ultrasonic wave emitter, a platen, an acoustic splitter and an acoustic detector. The invention may also include a method of scanning a fingerprint, wherein ultrasonic energy is provided and reflected with an acoustic splitter to provide first reflected energy, the first reflected energy may be reflected from a finger to provide second reflected energy, the second reflected energy may be received at the splitter, and allowed to pass through the splitter, and the second reflected energy may then be received at a detector.
Abstract: An image processing apparatus according to an aspect of the invention includes an LCD touch panel and a central control circuit. The LCD touch panel includes a display unit and an input unit. The central control circuit which is of an image processing unit performs a searching process for searching a portion having high similarity between a registered image and a processing target image and a registering process for producing and registering the registered image. The central control circuit deals with an input image as the processing target image when the searching process is to be performed, the central control circuit causes the display unit to display the input image when the registering process is performed, and the image processing unit draws an image used to produce the registered image unit based on operation of the input unit while overlapping the image with the input image displayed on the display.
Abstract: The present invention relates to a method of generating a keypoint descriptor for identifying an object in an image or in a sequence of images, the keypoint descriptor being substantially invariant to a transformation of the object in the image.
Type:
Grant
Filed:
November 5, 2009
Date of Patent:
April 29, 2014
Assignee:
Cortexica Vision Systems Limited
Inventors:
Anil Anthony Bharath, Jeffrey Ng Sing Kwong
Abstract: In a method of real-time 3-D visualization and navigation for interventional procedures a position is determined at which data will be acquired, and data is acquired at the position. Features are extracted from the acquired data. A model described by the extracted features is built or updated, wherein updating the model includes calculating at least one acquisition parameter of a set of acquisition parameters. 3-D scenes showing the interior or exterior of one or more organs are displayed using the model or data extracted from the model. A region for updating is determined.
Type:
Grant
Filed:
March 29, 2006
Date of Patent:
April 29, 2014
Assignee:
Siemens Medical Solutions USA, Inc.
Inventors:
Klaus J. Kirchberg, Christine H. Lorenz
Abstract: Provided are methods, computer implemented methods, and devices for machine vision. The methods and devices are capable of quantifying the amount of a particular color resulting from a particular stain in a sample stained with multiple agents.
Abstract: A system using face and iris image capture for recognition of people. The system may have wide field-of-view, medium field-of-view and narrow field-of-view cameras to capture images of a scene of people, faces and irises for processing and recognition. Matching of the face and iris images with images of a database may be a basis for recognition and identification of a subject person.
Type:
Grant
Filed:
March 2, 2007
Date of Patent:
April 22, 2014
Assignee:
Honeywell International Inc.
Inventors:
Gary E. Determan, Vincent C. Jacobson, Jan Jelinek, Thomas Phinney, Rida M. Hamza, Terry Ahrens, George A. Kilgore, Rand P. Whillock, Saad Bedros
Abstract: A method for evaluating quality of image representing a fingerprint pattern is provided, the method including performing a filtering process on a region of the image including a digital filtering of the selected region successively performed a certain number of times. According to the method, convergence of said filtering process is determined, and the image within the region is classified as an image with identifiable pattern if the filtering process is convergent, whereas the image within the region is classified as an image with unidentifiable pattern if the filtering process is nonconvergent. The digital filtering is performed using a filter adapted to at least one of the following local image parameters: curvature of dermal ridges, scale and tilt angle of dermal ridges.
Type:
Grant
Filed:
October 21, 2009
Date of Patent:
April 15, 2014
Inventors:
Vladimir Nickolaevich Bichigov, Pavel Anatolievich Zaytsev, Alexandr Vladimirovich Moksin, Ivan Borisovich Shapshal
Abstract: Systems and methods process an image having a plurality of pixels includes an image sensor to capture an image; a first-layer to encode local patches on an image region; and a second layer to jointly encode patches from the same image region.
Abstract: A disclosed identification method of identifying a data point distribution area on a coordinate plane includes dividing an area on the coordinate plane into divided areas so that the divided areas radiate from a division center point; selecting, in each of the divided areas, from among the data points in the divided area, a data point having the greatest distance from the division center point as a representative point; determining whether there is an overlapping area where a distribution representative point area overlaps a determination area; and determining, when there is the overlapping area, that the data group to be determined is a relevant data group.
Abstract: In the case where a three-dimensional image including a plurality of element figures, such as a quadrangular plane, a curved surface including a ridgeline, and a cylinder, is edited using an image editing apparatus such as a three-dimensional CAD apparatus, when one element figure in the three-dimensional image is specified, other element figures connected to the specified one element figure, or other element figures included in the same three-dimensional image in which the one element figure is included are retrieved, and other element figure in the retrieved other element figures, which was determined as having a specific relationship to the one element figure is related to the one element figure. Then, the specified element figure and the element figures related to the specified element figure are edited collectively based on an operation received by the image editing apparatus.
Abstract: A method for training a classifier to be operative as an epithelial texture classifier, includes obtaining a plurality of training micrograph areas of biopsy tissue and for each of the training micrograph areas, identifying probable locations of nuclei that form epithelia, generating a skeleton graph from the probable locations of the nuclei that form the epithelia, manually drawing walls on the skeleton graph outside of the epithelia to divide the epithelia from one another, and manually selecting points that lie entirely inside the epithelia to generate open and/or closed geodesic paths in the skeleton graph between pairs of the selected points. Data is obtained from points selected from the walls and the paths and applied to a classifier to train the classifier as the epithelial texture classifier.
Type:
Grant
Filed:
May 11, 2011
Date of Patent:
February 18, 2014
Assignees:
NEC Laboratories America, Inc., NEC Corporation
Inventors:
Christopher D. Malon, Atsushi Marugame, Eric Cosatto
Abstract: A disclosed method of identifying a data point distribution area having data points on a coordinate plane includes dividing a coordinate plane area into plural divided areas using lines; in each divided area, selecting outermost data point data as representative points, and connecting the selected representative points to define a distribution area; comparing the area with a reference area to determine an overlapping area; and determining that the distribution area is a relevant area based on the existence of the overlapping area.
Abstract: Various embodiments illustrated and described herein include at least one of systems, methods, and software that utilizes imagery from a single high-resolution camera to capture images, locate individuals, and provide images to a facial recognition process. Some embodiments also include prioritization and filtering that choose which captured images from a stream of images to process and when to process them when there are many images to be processed.
Abstract: A method and a device for adjusting an output frame are provided. The present disclosure is suitable for an electronic device with a display. In the present disclosure, a relative position between a user and the display is obtained first. Then, an output frame which will be outputted by the display later is adjusted according to the relative position. As a result, the output frame is adjusted in advance to fit the present position of the user, so that the user views the output frame more easily and enjoys the best display result.
Abstract: Systems and methods for retrieving, manipulating, and viewing 3D image objects from 3D virtual microscope slide images (“3D digital slides”) are provided. An image library module provides access to the imagery data in a 3D digital slide and constructs 3D image objects that are coextensive with the 3D digital slide or a 3D sub-portion thereof. From within the 3D image object, cross layer planar views spanning various depths of the 3D digital slide are constructed as well as 3D prisms and other shaped image areas. The image library module allows a 3D image object to be sliced into horizontal and vertical views, skewed cross layer views and regular and irregular shaped 3D image areas for viewing by a user.
Abstract: Training data object images are clustered as a function of motion direction attributes and resized from respective original into same aspect ratios. Motionlet detectors are learned for each of the sets from features extracted from the resized object blobs. A deformable sliding window is applied to detect an object blob in input by varying window size, shape or aspect ratio to conform to a shape of the detected input video object blob. A motion direction of an underlying image patch of the detected input video object blob is extracted and motionlet detectors selected and applied that have similar motion directions. An object is thus detected within the detected blob and semantic attributes of an underlying image patch extracted if a motionlet detectors fires, the extracted semantic attributes available for use for searching for the detected object.
Type:
Grant
Filed:
April 13, 2011
Date of Patent:
December 31, 2013
Assignee:
International Business Machines Corporation
Inventors:
Ankur Datta, Rogerio S. Feris, Sharathchandra U. Pankanti, Behjat Siddiquie, Yun Zhai
Abstract: A facial recognition method for eliminating the effect of noise blur and environmental variations is provided, applicable to a data processing apparatus, to determine whether a current-face-image matches a reference-face-image or not. According to the method, Gaussian Noise blur Reduction or another noise blur reduction method is performed on the current-face-image and the reference-face-image. Furthermore, the current-face-image and the reference-face-image are respectively divided into a plurality of blocks, so as to derive feature-vector sets representing the current-face-image and the reference-face-image. Finally a dynamic threshold value is determined according to environmental variations to be compared with the difference of the feature-vector sets, so as to determine whether a current-face-image matches a reference-face-image.