Patents by Inventor James William Philbin
James William Philbin 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: 10621420Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating numeric embeddings of images. One of the methods includes obtaining training images; generating a plurality of triplets of training images; and training a neural network on each of the triplets to determine trained values of a plurality of parameters of the neural network, wherein training the neural network comprises, for each of the triplets: processing the anchor image in the triplet using the neural network to generate a numeric embedding of the anchor image; processing the positive image in the triplet using the neural network to generate a numeric embedding of the positive image; processing the negative image in the triplet using the neural network to generate a numeric embedding of the negative image; computing a triplet loss; and adjusting the current values of the parameters of the neural network using the triplet loss.Type: GrantFiled: October 30, 2017Date of Patent: April 14, 2020Assignee: Google LLCInventors: James William Philbin, Gerhard Florian Schroff, Dmitry Kalenichenko
-
Publication number: 20180053042Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating numeric embeddings of images. One of the methods includes obtaining training images; generating a plurality of triplets of training images; and training a neural network on each of the triplets to determine trained values of a plurality of parameters of the neural network, wherein training the neural network comprises, for each of the triplets: processing the anchor image in the triplet using the neural network to generate a numeric embedding of the anchor image; processing the positive image in the triplet using the neural network to generate a numeric embedding of the positive image; processing the negative image in the triplet using the neural network to generate a numeric embedding of the negative image; computing a triplet loss; and adjusting the current values of the parameters of the neural network using the triplet loss.Type: ApplicationFiled: October 30, 2017Publication date: February 22, 2018Inventors: James William Philbin, Gerhard Florian Schroff, Dmitry Kalenichenko
-
Patent number: 9836641Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating numeric embeddings of images. One of the methods includes obtaining training images; generating a plurality of triplets of training images; and training a neural network on each of the triplets to determine trained values of a plurality of parameters of the neural network, wherein training the neural network comprises, for each of the triplets: processing the anchor image in the triplet using the neural network to generate a numeric embedding of the anchor image; processing the positive image in the triplet using the neural network to generate a numeric embedding of the positive image; processing the negative image in the triplet using the neural network to generate a numeric embedding of the negative image; computing a triplet loss; and adjusting the current values of the parameters of the neural network using the triplet loss.Type: GrantFiled: December 17, 2015Date of Patent: December 5, 2017Assignee: Google Inc.Inventors: James William Philbin, Gerhard Florian Schroff, Dmitry Kalenichenko
-
Patent number: 9442950Abstract: Systems and methods for a dynamic visual search engine are provided. In one example method, a criteria used to partition a set of compressed image descriptors into multiple database shards may be determined. Additionally, a size of a dynamic index may be determined. The dynamic index may represent a dynamic number of images and may be configured to accept insertion of reference images into the dynamic index that can be search against immediately. According to the method, an instruction to merge the uncompressed image descriptors of the dynamic index into the database shards of the compressed image descriptors may be received, and the uncompressed image descriptors of the dynamic index may be responsively merged into the database shards of the compressed image descriptors based on the criteria.Type: GrantFiled: October 29, 2014Date of Patent: September 13, 2016Assignee: Google Inc.Inventors: James William Philbin, Anand Pillai, John Flynn, Hartwig Adam
-
Publication number: 20160180151Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating numeric embeddings of images. One of the methods includes obtaining training images; generating a plurality of triplets of training images; and training a neural network on each of the triplets to determine trained values of a plurality of parameters of the neural network, wherein training the neural network comprises, for each of the triplets: processing the anchor image in the triplet using the neural network to generate a numeric embedding of the anchor image; processing the positive image in the triplet using the neural network to generate a numeric embedding of the positive image; processing the negative image in the triplet using the neural network to generate a numeric embedding of the negative image; computing a triplet loss; and adjusting the current values of the parameters of the neural network using the triplet loss.Type: ApplicationFiled: December 17, 2015Publication date: June 23, 2016Inventors: James William Philbin, Gerhard Florian Schroff, Dmitry Kalenichenko
-
Publication number: 20150169611Abstract: Systems and methods for a dynamic visual search engine are provided. In one example method, a criteria used to partition a set of compressed image descriptors into multiple database shards may be determined. Additionally, a size of a dynamic index may be determined. The dynamic index may represent a dynamic number of images and may be configured to accept insertion of reference images into the dynamic index that can be search against immediately. According to the method, an instruction to merge the uncompressed image descriptors of the dynamic index into the database shards of the compressed image descriptors may be received, and the uncompressed image descriptors of the dynamic index may be responsively merged into the database shards of the compressed image descriptors based on the criteria.Type: ApplicationFiled: October 29, 2014Publication date: June 18, 2015Inventors: James William Philbin, Anand Pillai, John Flynn, Hartwig Adam
-
Publication number: 20150169186Abstract: A method and apparatus for surfacing content during a photo sharing process is described. The method may include detecting the initiation of a social networking photo sharing process for a digital image. The method may also include transmitting digital image data corresponding to the digital image to a content surfacing server to locate content relevant to the digital image. The method may also include receiving one or more items of data relevant to the digital image from the content surfacing server. Furthermore, the method may include presenting a notification, before the digital image is uploaded to a social networking system by the social networking photo sharing process, to a user to indicate that content relevant to the digital image has been received.Type: ApplicationFiled: March 16, 2012Publication date: June 18, 2015Inventors: Hartmut Neven, Dragomir Anguelov, John Flynn, Mark Joseph Cummins, James William Philbin
-
Patent number: 8898139Abstract: Systems and methods for a dynamic visual search engine are provided. In one example method, a criteria used to partition a set of compressed image descriptors into multiple database shards may be determined. Additionally, a size of a dynamic index may be determined. The dynamic index may represent a dynamic number of images and may be configured to accept insertion of reference images into the dynamic index that can be search against immediately. According to the method, an instruction to merge the uncompressed image descriptors of the dynamic index into the database shards of the compressed image descriptors may be received, and the uncompressed image descriptors of the dynamic index may be responsively merged into the database shards of the compressed image descriptors based on the criteria.Type: GrantFiled: June 22, 2012Date of Patent: November 25, 2014Assignee: Google Inc.Inventors: James William Philbin, Anand Pillai, John Flynn, Hartwig Adam