Patents by Inventor Bernhard Schoelkopf
Bernhard Schoelkopf has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Patent number: 10032254Abstract: A computer-implemented method for recovering a digital image (x) from a sequence of observed digital images (y1, . . . , yT), includes: obtaining an observed digital image (yt); estimating a point spread function (ft) based on the observed image (yt); estimating the recovered digital image (x), based on the estimated point spread function (ft) and the observed image (yt); and repeating the above steps. In order to correct optical aberrations of a lens, a point spread function of the lens may be used.Type: GrantFiled: September 28, 2011Date of Patent: July 24, 2018Assignee: Max-Planck-Gesellschaft zur Förderung der Wissenschaften e.V.Inventors: Stefan Harmeling, Michael Hirsch, Suvrit Sra, Bernhard Schölkopf, Christian J. Schuler
-
Publication number: 20170212202Abstract: A method for correcting Bo fluctuation-induced ghosting artifacts in long-TE gradient-echo scan images, comprising the steps of: acquiring an image (u); determining phase offsets (?); and applying the phase offsets (?) to the image (u); such that an entropy of the spatial intensity variations in the corrected image (u) decreases.Type: ApplicationFiled: January 25, 2017Publication date: July 27, 2017Inventors: Alexander LOKTYUSHIN, Philipp EHSES, Klaus SCHEFFLER, Bernhard SCHÖLKOPF
-
Publication number: 20130242129Abstract: A computer-implemented method for recovering a digital image (x) from a sequence of observed digital images (y1, . . . , yT), includes: obtaining an observed digital image (yt); estimating a point spread function (ft) based on the observed image (yt); estimating the recovered digital image (x), based on the estimated point spread function (ft) and the observed image (yt); and repeating the above steps. In order to correct optical aberrations of a lens, a point spread function of the lens may be used.Type: ApplicationFiled: September 28, 2011Publication date: September 19, 2013Inventors: Stefan Harmeling, Michael Hirsch, Suvrit Sra, Bernhard Schölkopf, Christian J. Schuler
-
Patent number: 8290568Abstract: It is disclosed a system and method (12) for determining a property map (82) of an object, particularly a human being, based on at least a first image (84), particularly an magnetic resonance (MR) image, of the object. In the method (12), a structure of reference pairs is defined in a first step (96), wherein each reference pair (16-26) comprises at least two entries (62). The first entry represents a property value, particularly an attenuation value. The second entry (62) preferably represents a group of image points (67) belonging together, which is extracted particularly from MR images (28) and comprises an interesting image point corresponding to the property value. In another step (98) of the method (12) a plurality of training pairs (16-26) is provided. A structure of the training pairs (16-26) corresponds to the structure of reference pairs, and the entries of respective training pairs (16-26) are known.Type: GrantFiled: January 9, 2009Date of Patent: October 16, 2012Assignee: Eberhard-Karls-Universitat Tubingen UniversitatsklinikumInventors: Bernd Pichler, Matthias Hofmann, Bernhard Schölkopf, Florian Steinke
-
Patent number: 8209269Abstract: Learning machines, such as support vector machines, are used to analyze datasets to recognize patterns within the dataset using kernels that are selected according to the nature of the data to be analyzed. Where the datasets include an invariance transformation or noise, tangent vectors are defined to identify relationships between the invariance or noise and the training data points. A covariance matrix is formed using the tangent vectors, then used in generation of the kernel, which may be based on a kernel PCA map.Type: GrantFiled: August 25, 2010Date of Patent: June 26, 2012Assignee: Health Discovery CorporationInventors: Bernhard Schoelkopf, Olivier Chapelle
-
Patent number: 7970718Abstract: A group of features that has been identified as “significant” in being able to separate data into classes is evaluated using a support vector machine which separates the dataset into classes one feature at a time. After separation, an extremal margin value is assigned to each feature based on the distance between the lowest feature value in the first class and the highest feature value in the second class. Separately, extremal margin values are calculated for a normal distribution within a large number of randomly drawn example sets for the two classes to determine the number of examples within the normal distribution that would have a specified extremal margin value. Using p-values calculated for the normal distribution, a desired p-value is selected. The specified extremal margin value corresponding to the selected p-value is compared to the calculated extremal margin values for the group of features.Type: GrantFiled: September 26, 2010Date of Patent: June 28, 2011Assignee: Health Discovery CorporationInventors: Isabelle Guyon, Andre Elisseeff, Bernhard Schoelkopf, Jason Aaron Edward Weston, Fernando Perez-Cruz
-
Publication number: 20110119213Abstract: Identification of a determinative subset of features from within a group of features is performed by training a support vector machine using training samples with class labels to determine a value of each feature, where features are removed based on their the value. One or more features having the smallest values are removed and an updated kernel matrix is generated using the remaining features. The process is repeated until a predetermined number of features remain which are capable of accurately separating the data into different classes.Type: ApplicationFiled: December 1, 2010Publication date: May 19, 2011Applicant: HEALTH DISCOVERY CORPORATIONInventors: André Elisseeff, Bernhard Schölkopf, Fernando Perez-Cruz
-
Publication number: 20110106735Abstract: Identification of a determinative subset of features from within a group of features is performed by training a support vector machine using training samples with class labels to determine a value of each feature, where features are removed based on their the value. One or more features having the smallest values are removed and an updated kernel matrix is generated using the remaining features. The process is repeated until a predetermined number of features remain which are capable of accurately separating the data into different classes. In some embodiments, features are eliminated by a ranking criterion based on a Lagrange multiplier corresponding to each training sample.Type: ApplicationFiled: November 11, 2010Publication date: May 5, 2011Applicant: HEALTH DISCOVERY CORPORATIONInventors: Jason Weston, André Elisseeff, Bernhard Schölkopf, Fernando Perez-Cruz, Isabelle Guyon
-
Publication number: 20110078099Abstract: A group of features that has been identified as “significant” in being able to separate data into classes is evaluated using a support vector machine which separates the dataset into classes one feature at a time. After separation, an extremal margin value is assigned to each feature based on the distance between the lowest feature value in the first class and the highest feature value in the second class. Separately, extremal margin values are calculated for a normal distribution within a large number of randomly drawn example sets for the two classes to determine the number of examples within the normal distribution that would have a specified extremal margin value. Using p-values calculated for the normal distribution, a desired p-value is selected. The specified extremal margin value corresponding to the selected p-value is compared to the calculated extremal margin values for the group of features.Type: ApplicationFiled: September 26, 2010Publication date: March 31, 2011Applicant: HEALTH DISCOVERY CORPORATIONInventors: Jason Aaron Edward Weston, André Elisseeff, Bernhard Schöelkopf, Fernando Perez-Cruz, Isabelle Guyon
-
Publication number: 20100318482Abstract: Learning machines, such as support vector machines, are used to analyze datasets to recognize patterns within the dataset using kernels that are selected according to the nature of the data to be analyzed. Where the datasets include an invariance transformation or noise, tangent vectors are defined to identify relationships between the invariance or noise and the training data points. A covariance matrix is formed using the tangent vectors, then used in generation of the kernel, which may be based on a kernel PCA map.Type: ApplicationFiled: August 25, 2010Publication date: December 16, 2010Applicant: HEALTH DISCOVERY CORPORATIONInventors: Peter L. Bartlett, André Elisseeff, Bernhard Schoelkopf, Olivier Chapelle
-
Patent number: 7805388Abstract: In a pre-processing step prior to training a learning machine, pre-processing includes reducing the quantity of features to be processed using feature selection methods selected from the group consisting of recursive feature elimination (RFE), minimizing the number of non-zero parameters of the system (l0-norm minimization), evaluation of cost function to identify a subset of features that are compatible with constraints imposed by the learning set, unbalanced correlation score, transductive feature selection and single feature using margin-based ranking. The features remaining after feature selection are then used to train a learning machine for purposes of pattern classification, regression, clustering and/or novelty detection.Type: GrantFiled: October 30, 2007Date of Patent: September 28, 2010Assignee: Health Discovery CorporationInventors: Jason Weston, André Elisseeff, Bernhard Schölkopf, Fernando Perez-Cruz, Isabelle Guyon
-
Patent number: 7788193Abstract: Learning machines, such as support vector machines, are used to analyze datasets to recognize patterns within the dataset using kernels that are selected according to the nature of the data to be analyzed. Where the datasets possesses structural characteristics, locational kernels can be utilized to provide measures of similarity among data points within the dataset. The locational kernels are then combined to generate a decision function, or kernel, that can be used to analyze the dataset. Where an invariance transformation or noise is present, tangent vectors are defined to identify relationships between the invariance or noise and the data points. A covariance matrix is formed using the tangent vectors, then used in generation of the kernel.Type: GrantFiled: October 30, 2007Date of Patent: August 31, 2010Assignee: Health Discovery CorporationInventors: Peter L. Bartlett, André Elisseeff, Bernhard Schoelkopf, Olivier Chapelle
-
Patent number: 7565409Abstract: What is disclosed is acquiring information regarding a web page, without having to commit to downloading that page. In one embodiment, after a current web page is downloaded from one source, and information regarding web pages linked to by links in the current web page are downloaded from a second source, when a user hovers a cursor over a link on a current web page, an informational region is displayed by the link that includes the information from the second source. The informational region may include, for example, a text box that apparently floats by the link. The information in the region can include, for example, keywords in the meta tags of the web page; paragraph headings of the web page; links on the web page to other pages; etc.Type: GrantFiled: November 28, 2006Date of Patent: July 21, 2009Assignee: Microsoft CorporationInventors: Lisa Heilbron, John C. Platt, Patrice Y. Simard, Bernhard Schoelkopf
-
Patent number: 7475048Abstract: A computer-implemented method is provided for ranking features within a large dataset containing a large number of features according to each feature's ability to separate data into classes. For each feature, a support vector machine separates the dataset into two classes and determines the margins between extremal points in the two classes. The margins for all of the features are compared and the features are ranked based upon the size of the margin, with the highest ranked features corresponding to the largest margins. A subset of features for classifying the dataset is selected from a group of the highest ranked features. In one embodiment, the method is used to identify the best genes for disease prediction and diagnosis using gene expression data from micro-arrays.Type: GrantFiled: November 7, 2002Date of Patent: January 6, 2009Assignee: Health Discovery CorporationInventors: Jason Weston, André Elisseeff, Bernhard Schölkopf, Fernando Perez-Cruz, Isabelle Guyon
-
Publication number: 20080301070Abstract: Learning machines, such as support vector machines, are used to analyze datasets to recognize patterns within the dataset using kernels that are selected according to the nature of the data to be analyzed. Where the datasets possesses structural characteristics, locational kernels can be utilized to provide measures of similarity among data points within the dataset. The locational kernels are then combined to generate a decision function, or kernel, that can be used to analyze the dataset. Where an invariance transformation or noise is present, tangent vectors are defined to identify relationships between the invariance or noise and the data points. A covariance matrix is formed using the tangent vectors, then used in generation of the kernel.Type: ApplicationFiled: October 30, 2007Publication date: December 4, 2008Inventors: Peter L. Bartlett, Andre Elisseeff, Bernhard Schoelkopf, Olivier Chapelle
-
Patent number: 7444015Abstract: A system and process for creating an interactive digital image, which allows a viewer to interact with a displayed image so as to change it with regard to a desired effect, such as exposure, focus or color, among others. An interactive image includes representative images which depict a scene with some image parameter varying between them. The interactive image also includes an index image, whose pixels each identify the representative image that exhibits the desired effect related to the varied image parameter at a corresponding pixel location. For example, a pixel of the index image might identify the representative image having a correspondingly-located pixel that depicts a portion of the scene at the sharpest focus. One primary form of interaction involves selecting a pixel of a displayed image whereupon the representative image identified in the index image at a corresponding pixel location is displayed in lieu of the currently displayed image.Type: GrantFiled: December 22, 2004Date of Patent: October 28, 2008Assignee: Microsoft CorporationInventors: Bernhard Schoelkopf, Kentaro Toyama, Matthew Uyttendaele
-
Patent number: 7444016Abstract: A system and process for creating an interactive digital image, which allows a viewer to interact with a displayed image so as to change it with regard to a desired effect, such as exposure, focus or color, among others. An interactive image includes representative images which depict a scene with some image parameter varying between them. The interactive image also includes an index image, whose pixels each identify the representative image that exhibits the desired effect related to the varied image parameter at a corresponding pixel location. For example, a pixel of the index image might identify the representative image having a correspondingly-located pixel that depicts a portion of the scene at the sharpest focus. One primary form of interaction involves selecting a pixel of a displayed image whereupon the representative image identified in the index image at a corresponding pixel location is displayed in lieu of the currently displayed image.Type: GrantFiled: November 9, 2005Date of Patent: October 28, 2008Assignee: Microsoft CorporationInventors: Bernhard Schoelkopf, Kentaro Toyama, Matthew Uyttendaele
-
Patent number: 7421115Abstract: A system and process for creating an interactive digital image, which allows a viewer to interact with a displayed image so as to change it with regard to a desired effect, such as exposure, focus or color, among others. An interactive image includes representative images which depict a scene with some image parameter varying between them. The interactive image also includes an index image, whose pixels each identify the representative image that exhibits the desired effect related to the varied image parameter at a corresponding pixel location. For example, a pixel of the index image might identify the representative image having a correspondingly-located pixel that depicts a portion of the scene at the sharpest focus. One primary form of interaction involves selecting a pixel of a displayed image whereupon the representative image identified in the index image at a corresponding pixel location is displayed in lieu of the currently displayed image.Type: GrantFiled: December 22, 2004Date of Patent: September 2, 2008Assignee: Microsoft CorporationInventors: Bernhard Schoelkopf, Kentaro Toyama, Matthew Uyttendaele
-
Patent number: 7391908Abstract: Systems and methods for object or pattern detection that use a nonlinear support vector (SV) machine are described. In the illustrated and described embodiment, objects or patterns comprising faces are detected. The decision surface is approximated in terms of a reduced set of expansion vectors. In order to determine the presence of a face, the kernelized inner product of the expansion vectors with the input pattern are sequentially evaluated and summed, such that if at any point the pattern can be rejected as not comprising a face, no more expansion vectors are used. The sequential application of the expansion vectors produces a substantial saving in computational time.Type: GrantFiled: February 28, 2005Date of Patent: June 24, 2008Assignee: Microsoft CorporationInventors: Andrew Blake, Sami Romdhani, Bernhard Schoelkopf, Philip H. S. Torr
-
Patent number: 7353215Abstract: Learning machines, such as support vector machines, are used to analyze datasets to recognize patterns within the dataset using kernels that are selected according to the nature of the data to be analyzed. Where the datasets possesses structural characteristics, locational kernels can be utilized to provide measures of similarity among data points within the dataset. The locational kernels are then combined to generate a decision function, or kernel, that can be used to analyze the dataset. Where invariance transformations or noise is present, tangent vectors are defined to identify relationships between the invariance or noise and the data points. A covariance matrix is formed using the tangent vectors, then used in generation of the kernel for recognizing patterns in the dataset.Type: GrantFiled: May 7, 2002Date of Patent: April 1, 2008Assignee: Health Discovery CorporationInventors: Peter L. Bartlett, André Elisseeff, Bernhard Schoelkopf