Patents by Inventor Ramin Zabih
Ramin Zabih 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: 11115678Abstract: An apparatus for encoding a current frame of a video. The apparatus includes a memory and a processor. The processor is configured to execute instructions stored in the memory to generate, for each reference frame of a subset of available reference frames, at least one respective candidate global motion model (GMM); partition the current frame into blocks; generate an aggregated residual frame for the current frame; and encode the respective residual blocks in a compressed bitstream. To generate the aggregated residual frame includes to select, for predicting each block of the blocks, a respective selected GMM, where the respective selected GMM corresponds to the one of the at least one respective candidate GMMs that minimizes a total error associated with the aggregated residual frame; and obtain respective residual blocks for the block.Type: GrantFiled: April 29, 2020Date of Patent: September 7, 2021Assignee: GOOGLE LLCInventors: Debargha Mukherjee, Emil Keyder, Michele Covell, Chen Wang, Sarah Parker, Ramin Zabih
-
Publication number: 20200260112Abstract: An apparatus for encoding a current frame of a video. The apparatus includes a memory and a processor. The processor is configured to execute instructions stored in the memory to generate, for each reference frame of a subset of available reference frames, at least one respective candidate global motion model (GMM); partition the current frame into blocks; generate an aggregated residual frame for the current frame; and encode the respective residual blocks in a compressed bitstream. To generate the aggregated residual frame includes to select, for predicting each block of the blocks, a respective selected GMM, where the respective selected GMM corresponds to the one of the at least one respective candidate GMMs that minimizes a total error associated with the aggregated residual frame; and obtain respective residual blocks for the block.Type: ApplicationFiled: April 29, 2020Publication date: August 13, 2020Inventors: Debargha Mukherjee, Emil Keyder, Michele Covell, Chen Wang, Sarah Parker, Ramin Zabih
-
Patent number: 10681374Abstract: A method for encoding a current frame of a video includes jointly determining respective motion models for reference frames and encoding the current frame using the respective motion models. The reference frames are used for encoding the current frame. Jointly determining respective motion models for reference frames includes determining respective aggregated residuals for combinations of candidate motion models and selecting the combination of candidate motion models that corresponds to the smallest aggregated residual. The respective motion models correspond to the candidate motion models of the selected combination.Type: GrantFiled: June 25, 2018Date of Patent: June 9, 2020Assignee: GOOGLE LLCInventors: Debargha Mukherjee, Emil Keyder, Michele Covell, Chen Wang, Sarah Parker, Ramin Zabih
-
Publication number: 20190149841Abstract: A method for encoding a current frame of a video includes jointly determining respective motion models for reference frames and encoding the current frame using the respective motion models. The reference frames are used for encoding the current frame. Jointly determining respective motion models for reference frames includes determining respective aggregated residuals for combinations of candidate motion models and selecting the combination of candidate motion models that corresponds to the smallest aggregated residual. The respective motion models correspond to the candidate motion models of the selected combination.Type: ApplicationFiled: June 25, 2018Publication date: May 16, 2019Inventors: Debargha Mukherjee, Emil Keyder, Michele Covell, Chen Wang, Sarah Parker, Ramin Zabih
-
Publication number: 20120106817Abstract: Systems and method can be employed to automatically extract radiation dose information from medical images, particularly a plurality of heterogeneous CT images including those from legacy CT scanners. A report including the extracted radiation dose information can be generated, and alerts can be sent to reduce the possibility of overexposures.Type: ApplicationFiled: April 29, 2011Publication date: May 3, 2012Inventors: George Shih, Devin Kenney, Ramin Zabih, Edward L. Nickoloff, Zheng Feng Lu
-
Patent number: 7734078Abstract: In time-resolved contrast-enhanced magnetic resonance angiography, a measure quantifying image quality provides a basis for generating a linear filtered composite image by facilitating selection of a mask and an arterial phase image for subtraction. Filtering of individual pixels of a temporal series of images provides enhanced contrast in a single image by allowing the temporal behavior of the pixel intensity to denote representation as an artery, vein or background tissue. Motion artifacts are suppressed by re-registering sequential images, adjusting weighting before averaging and subtraction and filtering the Fourier data to eliminate data corrupted by motion or other phenomena.Type: GrantFiled: June 1, 2009Date of Patent: June 8, 2010Assignee: Cornell Research Foundation, Inc.Inventors: Martin R. Prince, Ramin Zabih, Yi Wang, Richard Watts, Junhwan Kim, Jeff Bezanson, Hale Ersoy
-
Publication number: 20090245606Abstract: In time-resolved contrast-enhanced magnetic resonance angiography, a measure quantifying image quality provides a basis for generating a linear filtered composite image by facilitating selection of a mask and an arterial phase image for subtraction. Filtering of individual pixels of a temporal series of images provides enhanced contrast in a single image by allowing the temporal behavior of the pixel intensity to denote representation as an artery, vein or background tissue. Motion artifacts are suppressed by re-registering sequential images, adjusting weighting before averaging and subtraction and filtering the Fourier data to eliminate data corrupted by motion or other phenomena.Type: ApplicationFiled: June 1, 2009Publication date: October 1, 2009Applicant: Cornell Research Foundation, Inc.Inventors: Martin R. Prince, Ramin Zabih, Yi Wang, Richard Watts, Junhwan Kim, Jeff Bezanson, Hale Ersoy
-
Patent number: 7545967Abstract: In time-resolved contrast-enhanced magnetic resonance angiography, a measure quantifying image quality provides a basis for generating a linear filtered composite image by facilitating selection of a mask and an arterial phase image for subtraction. Filtering of individual pixels of a temporal series of images provides enhanced contrast in a single image by allowing the temporal behavior of the pixel intensity to denote representation as an artery, vein or background tissue. Motion artifacts are suppressed by re-registering sequential images, adjusting weighting before averaging and subtraction and filtering the Fourier data to eliminate data corrupted by motion or other phenomena.Type: GrantFiled: September 17, 2003Date of Patent: June 9, 2009Assignee: Cornell Research Foundation Inc.Inventors: Martin R. Prince, Ramin Zabih, Yi Wang, Richard Watts, Junhwan Kim, Jeff Bezanson, Hale Ersoy
-
Publication number: 20080080775Abstract: Methods and systems for reconstructing an object from observations of interaction of the object with a physical system.Type: ApplicationFiled: September 28, 2007Publication date: April 3, 2008Applicant: Cornell Center for Technology Enterprise & CommercializationInventors: Ramin Zabih, Ashish Raj
-
Patent number: 7123757Abstract: An efficient navigator method is presented that substantially increases the scan efficiency while maintaining the motion suppression effectiveness in magnetic resonance imaging. The method is achieved by simultaneously acquiring different image volumes at different motion states of the subject being scanned. A scheduling algorithm is used to assign volumes to position bins of a motion histogram of the subject. The motion histogram is periodically updated and the volumes are reassigned to position bins.Type: GrantFiled: June 6, 2002Date of Patent: October 17, 2006Assignee: Cornell Research Foundation, Inc.Inventors: Vladimir N. Kolmogorov, Yi Wang, Ramin Zabih
-
Publication number: 20020188191Abstract: An efficient navigator method is presented that substantially increases the scan efficiency while maintaining the motion suppression effectiveness in magnetic resonance imaging. The method is achieved by simultaneously acquiring different image volumes at different motion states of the subject being scanned. A scheduling algorithm is used to assign volumes to position bins of a motion histogram of the subject. The motion histogram is periodically updated and the volumes are reassigned to position bins.Type: ApplicationFiled: June 6, 2002Publication date: December 12, 2002Applicant: Cornell Research Foundation, Inc.Inventors: Vladimir N. Kolmogorov, Yi Wang, Ramin Zabih
-
Patent number: 6226407Abstract: An invention for quantitatively analyzing bitmap images using computer vision is disclosed. In one embodiment, these bitmap images correspond to the rendered screens of two applications based on a same data source file. An optional preprocessing step filters non-important information for analyzing purposes such as the background, replaces each image with a blank image of the same size, and modifies non-visible attributes of a screen so their structure can be recovered during the computer vision processing. After a series of rendered screens of each application are captured, these bitmap images are interpreted using computer vision techniques to produce machine-readable visual attributes of the rendered screens. Corresponding attributes from each of the two applications are then compared to generate a set of differences, where those differences within a predefined set of deliberate design differences are ignored.Type: GrantFiled: March 18, 1998Date of Patent: May 1, 2001Assignee: Microsoft CorporationInventors: Ramin Zabih, Ramin Halviatti
-
Patent number: 5767922Abstract: Apparatus and process for detecting scene breaks in a sequence of video frames providing a moving picture. Entering and exiting edge pixels in each of a plurality of successive video frames are counted, and an edge change fraction for each of the successive video frames is derived therefrom. Peaks which are detected in the edge change fractions are indicative of scene breaks.Type: GrantFiled: April 5, 1996Date of Patent: June 16, 1998Assignee: Cornell Research Foundation, Inc.Inventors: Ramin Zabih, Justin F. Miller, Kevin W. Mai