Patents by Inventor Ara Danielyan
Ara Danielyan 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: 11928773Abstract: In various embodiments, a training application generates a trained encoder that automatically generates shape embeddings having a first size and representing three-dimensional (3D) geometry shapes. First, the training application generates a different view activation for each of multiple views associated with a first 3D geometry based on a first convolutional neural network (CNN) block. The training application then aggregates the view activations to generate a tiled activation. Subsequently, the training application generates a first shape embedding having the first size based on the tiled activation and a second CNN block. The training application then generates multiple re-constructed views based on the first shape embedding. The training application performs training operation(s) on at least one of the first CNN block and the second CNN block based on the views and the re-constructed views to generate the trained encoder.Type: GrantFiled: February 23, 2022Date of Patent: March 12, 2024Assignee: AUTODESK, INC.Inventors: Thomas Ryan Davies, Michael Haley, Ara Danielyan, Morgan Fabian
-
Patent number: 11741662Abstract: In various embodiments, a training application generates a trained encoder that automatically generates shape embeddings having a first size and representing three-dimensional (3D) geometry shapes. First, the training application generates a different view activation for each of multiple views associated with a first 3D geometry based on a first convolutional neural network (CNN) block. The training application then aggregates the view activations to generate a tiled activation. Subsequently, the training application generates a first shape embedding having the first size based on the tiled activation and a second CNN block. The training application then generates multiple re-constructed views based on the first shape embedding. The training application performs training operation(s) on at least one of the first CNN block and the second CNN block based on the views and the re-constructed views to generate the trained encoder.Type: GrantFiled: October 29, 2018Date of Patent: August 29, 2023Assignee: AUTODESK, INC.Inventors: Thomas Davies, Michael Haley, Ara Danielyan, Morgan Fabian
-
Patent number: 11380045Abstract: In various embodiments, a training application generates a trained encoder that automatically generates shape embeddings having a first size and representing three-dimensional (3D) geometry shapes. First, the training application generates a different view activation for each of multiple views associated with a first 3D geometry based on a first convolutional neural network (CNN) block. The training application then aggregates the view activations to generate a tiled activation. Subsequently, the training application generates a first shape embedding having the first size based on the tiled activation and a second CNN block. The training application then generates multiple re-constructed views based on the first shape embedding. The training application performs training operation(s) on at least one of the first CNN block and the second CNN block based on the views and the re-constructed views to generate the trained encoder.Type: GrantFiled: October 29, 2018Date of Patent: July 5, 2022Assignee: AUTODESK, INC.Inventors: Thomas Davies, Michael Haley, Ara Danielyan, Morgan Fabian
-
Publication number: 20220180596Abstract: In various embodiments, a training application generates a trained encoder that automatically generates shape embeddings having a first size and representing three-dimensional (3D) geometry shapes. First, the training application generates a different view activation for each of multiple views associated with a first 3D geometry based on a first convolutional neural network (CNN) block. The training application then aggregates the view activations to generate a tiled activation. Subsequently, the training application generates a first shape embedding having the first size based on the tiled activation and a second CNN block. The training application then generates multiple re-constructed views based on the first shape embedding. The training application performs training operation(s) on at least one of the first CNN block and the second CNN block based on the views and the re-constructed views to generate the trained encoder.Type: ApplicationFiled: February 23, 2022Publication date: June 9, 2022Inventors: Thomas Ryan DAVIES, Michael HALEY, Ara DANIELYAN, Morgan FABIAN
-
Patent number: 11126330Abstract: In various embodiments, a training application generates a trained encoder that automatically generates shape embeddings having a first size and representing three-dimensional (3D) geometry shapes, First, the training application generates a different view activation for each of multiple views associated with a first 3D geometry based on a first convolutional neural network (CNN) block. The training application then aggregates the view activations to generate a tiled activation. Subsequently, the training application generates a first shape embedding having the first size based on the tiled activation and a second CNN block. The training application then generates multiple re-constructed views based on the first shape embedding. The training application performs training operation(s) on at least one of the first CNN block and the second CNN block based on the views and the re-constructed views to generate the trained encoder.Type: GrantFiled: October 29, 2018Date of Patent: September 21, 2021Assignee: AUTODESK, INC.Inventors: Thomas Davies, Michael Haley, Ara Danielyan, Morgan Fabian
-
Patent number: 10909307Abstract: A system and technique for capturing a workflow history and video of an electronic document are disclosed. Events generated by an application while modifying an electronic document are stored on a web server as metadata. In addition, a captured digital image or frames of captured digital video that reflect the state of the document at the time the event was generated are also stored on the web server. The metadata is associated with one or more portions of the document and with the captured digital image or frames of captured digital video.Type: GrantFiled: November 27, 2012Date of Patent: February 2, 2021Assignee: AUTODESK, INC.Inventors: Tovi Grossman, George Fitzmaurice, Justin Frank Matejka, Thomas White, Ara Danielyan, Ruslana Steininger, Michael Chen, Anderson Nogueira
-
Publication number: 20200134908Abstract: In various embodiments, a training application generates a trained encoder that automatically generates shape embeddings having a first size and representing three-dimensional (3D) geometry shapes. First, the training application generates a different view activation for each of multiple views associated with a first 3D geometry based on a first convolutional neural network (CNN) block. The training application then aggregates the view activations to generate a tiled activation. Subsequently, the training application generates a first shape embedding having the first size based on the tiled activation and a second CNN block. The training application then generates multiple re-constructed views based on the first shape embedding. The training application performs training operation(s) on at least one of the first CNN block and the second CNN block based on the views and the re-constructed views to generate the trained encoder.Type: ApplicationFiled: October 29, 2018Publication date: April 30, 2020Inventors: Thomas DAVIES, Michael HALEY, Ara DANIELYAN, Morgan FABIAN
-
Publication number: 20200134909Abstract: In various embodiments, a training application generates a trained encoder that automatically generates shape embeddings having a first size and representing three-dimensional (3D) geometry shapes. First, the training application generates a different view activation for each of multiple views associated with a first 3D geometry based on a first convolutional neural network (CNN) block. The training application then aggregates the view activations to generate a tiled activation. Subsequently, the training application generates a first shape embedding having the first size based on the tiled activation and a second CNN block. The training application then generates multiple re-constructed views based on the first shape embedding. The training application performs training operation(s) on at least one of the first CNN block and the second CNN block based on the views and the re-constructed views to generate the trained encoder.Type: ApplicationFiled: October 29, 2018Publication date: April 30, 2020Inventors: Thomas DAVIES, Michael HALEY, Ara DANIELYAN, Morgan FABIAN
-
Publication number: 20200133449Abstract: In various embodiments, a training application generates a trained encoder that automatically generates shape embeddings having a first size and representing three-dimensional (3D) geometry shapes, First, the training application generates a different view activation for each of multiple views associated with a first 3D geometry based on a first convolutional neural network (CNN) block. The training application then aggregates the view activations to generate a tiled activation. Subsequently, the training application generates a first shape embedding having the first size based on the tiled activation and a second CNN block. The training application then generates multiple re-constructed views based on the first shape embedding. The training application performs training operation(s) on at least one of the first CNN block and the second CNN block based on the views and the re-constructed views to generate the trained encoder.Type: ApplicationFiled: October 29, 2018Publication date: April 30, 2020Inventors: Thomas DAVIES, Michael HALEY, Ara DANIELYAN, Morgan FABIAN
-
Patent number: 10515143Abstract: A system and technique for capturing a workflow history and video of an electronic document are disclosed. Events generated by an application while modifying an electronic document are stored on a web server as metadata. In addition, a captured digital image or frames of captured digital video that reflect the state of the document at the time the event was generated are also stored on the web server. The metadata is associated with one or more portions of the document and with the captured digital image or frames of captured digital video.Type: GrantFiled: December 5, 2012Date of Patent: December 24, 2019Assignee: AUTODESK, INC.Inventors: Tovi Grossman, George Fitzmaurice, Justin Frank Matejka, Thomas White, Ara Danielyan, Ruslana Steininger, Michael Chen, Anderson Nogueira