-
公开(公告)号：GB2626398B
公开(公告)日：2025-02-19
申请号：GB202317072
申请日：2023-11-07
Applicant: ADOBE INC
Inventor： JONATHAN BRANDT , SCOTT COHEN , ZHE LIN , ZHIHONG DING , DARSHAN G PRASAD , MATTHEW JOSS , CELSO GOMES , JIANMING ZHANG , OLENA SOROKA , KLAAS STOECKMANN , MICHAEL ZIMMERMANN , THOMAS MUEHRKE
Abstract: The present disclosure relates to systems, methods, and non-transitory computer-readable media that modify digital images via scene-based editing using image understanding facilitated by artificial intelligence. For instance, in one or more embodiments, the disclosed systems generate utilizing a segmentation neural network, an object mask for each object of a plurality of objects of a digital image. The disclosed systems detect a first user interaction with an object in the digital image displayed via a graphical user interface. The disclosed systems surface, via the graphical user interface, the object mask for the object in response to the first user interaction. The disclosed systems perform an object-aware modification of the digital image in response to a second user interaction with the object mask for the object.
-
公开(公告)号：GB2620467B
公开(公告)日：2024-10-02
申请号：GB202305010
申请日：2023-04-04
Applicant: ADOBE INC
Inventor： ZHIFEI ZHANG , ZHE LIN , SCOTT COHEN , DARSHAN G PRASAD , ZHIHONG DING
Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods for transferring global style features between digital images utilizing one or more machine learning models or neural networks. In particular, in one or more embodiments, the disclosed systems receive a request to transfer a global style from a source digital image to a target digital image, identify at least one target object within the target digital image, and transfer the global style from the source digital image to the target digital image while maintaining an object style of the at least one target object.
-
公开(公告)号：GB2620467A
公开(公告)日：2024-01-10
申请号：GB202305010
申请日：2023-04-04
Applicant: ADOBE INC
Inventor： ZHIFEI ZHANG , ZHE LIN , SCOTT COHEN , DARSHAN G PRASAD , ZHIHONG DING
Abstract: Upon receipt of a request to transfer a global style from a source digital image to a target digital image, at least one target object is identified within the target image, and the global style is transferred from from the source image to the target image while the style of the target object(s) is maintained. This is achieved by removing the object(s) from the target image, applying a style transfer neural network to transfer the style to a version of the target image which has the object(s) removed, then re-inserting the object(s). The target objects may be identified using a machine learning object detection model, the target object(s) may be foreground object(s), maybe harmonised with the background before re-insertion, and their extraction may be achieved utilising a segmentation model. Object(s) may also be extracted from the source image, producing an intermediate image for defining the global style for transfer. The neural networks may be encoder neural networks for extracting global code representing features corresponding to overall appearance from the source image and spatial code representing geometric layout from the target image and generator neural networks for combining global and spatial codes to generate a modified digital image.
-
公开(公告)号：GB2572234A
公开(公告)日：2019-09-25
申请号：GB201817945
申请日：2018-11-02
Applicant: ADOBE INC
Inventor： WALTER WEI-TUH CHANG , ZHIHONG DING , LUBOMIRA A DONTCHEVA , GREGG DARRYL WILENSKY , DARSHAN G PRASAD , CLAUDIA VERONICA ROBERTS
Abstract: A computer program and associated method and system for integrating a tutorial into an application wherein a tutorial file derived from a web page is parsed, the tutorial file being constructed in a hierarchical format in which tutorial attributes are specified with respect to corresponding values which include at least one instruction included in the web page for a software application 302. A mapping is executed between the attributes and application features of the software application 304, and instructional code or a data file is generated that is executable by a native instruction service of the software application to generate at least one instruction element within the application based on the mapping identifying at least one application feature which would enable completion of an instruction 306. The hierarchical format may have a number of tutorial steps and each step may have an identifier indicating an application feature. Executing the mapping may utilise natural language processing of the tutorial attributes to relate them to application features.
-
-
-
Search Results
4
records
(took 0.019 seconds)
