-
公开(公告)号：GB2623401B
公开(公告)日：2025-01-08
申请号：GB202311866
申请日：2023-08-02
Applicant: ADOBE INC
Inventor： ZHE LIN , HAITIAN ZHENG , ELYA SHECHTMAN , JIANMING ZHANG , JINGWAN LU , NING XU , QING LIU , SCOTT COHEN , SOHRAB AMIRGHODSI
Abstract: The present disclosure relates to systems, methods, and non-transitory computer readable media for panoptically guiding digital image inpainting utilizing a panoptic inpainting neural network. In some embodiments, the disclosed systems utilize a panoptic inpainting neural network to generate an inpainted digital image according to panoptic segmentation map that defines pixel regions corresponding to different panoptic labels. In some cases, the disclosed systems train a neural network utilizing a semantic discriminator that facilitates generation of digital images that are realistic while also conforming to a semantic segmentation. The disclosed systems generate and provide a panoptic inpainting interface to facilitate user interaction for inpainting digital images. In certain embodiments, the disclosed systems iteratively update an inpainted digital image based on changes to a panoptic segmentation map.
-
2.发明专利
公开(公告)号：GB2623402B
公开(公告)日：2025-04-23
申请号：GB202311871
申请日：2023-08-02
Applicant: ADOBE INC
Inventor： ZHE LIN , HAITIAN ZHENG , ELYA SHECHTMAN , JIANMING ZHANG , JINGWAN LU , NING XU , QING LIU , SCOTT COHEN , SOHRAB AMIRGHODSI
Abstract: The present disclosure relates to systems, methods, and non-transitory computer readable media for panoptically guiding digital image inpainting utilizing a panoptic inpainting neural network. In some embodiments, the disclosed systems utilize a panoptic inpainting neural network to generate an inpainted digital image according to panoptic segmentation map that defines pixel regions corresponding to different panoptic labels. In some cases, the disclosed systems train a neural network utilizing a semantic discriminator that facilitates generation of digital images that are realistic while also conforming to a semantic segmentation. The disclosed systems generate and provide a panoptic inpainting interface to facilitate user interaction for inpainting digital images. In certain embodiments, the disclosed systems iteratively update an inpainted digital image based on changes to a panoptic segmentation map.
-
公开(公告)号：GB2623162B
公开(公告)日：2025-01-01
申请号：GB202311936
申请日：2023-08-03
Applicant: ADOBE INC
Inventor： ZHE LIN , HAITIAN ZHENG , ELYA SHECHTMAN , JIANMING ZHANG , JINGWAN LU , NING XU , QING LIU , SCOTT COHEN , SOHRAB AMIRGHODSI
Abstract: The present disclosure relates to systems, methods, and non-transitory computer readable media for panoptically guiding digital image inpainting utilizing a panoptic inpainting neural network. In some embodiments, the disclosed systems utilize a panoptic inpainting neural network to generate an inpainted digital image according to panoptic segmentation map that defines pixel regions corresponding to different panoptic labels. In some cases, the disclosed systems train a neural network utilizing a semantic discriminator that facilitates generation of digital images that are realistic while also conforming to a semantic segmentation. The disclosed systems generate and provide a panoptic inpainting interface to facilitate user interaction for inpainting digital images. In certain embodiments, the disclosed systems iteratively update an inpainted digital image based on changes to a panoptic segmentation map.
-
Patent Agency Ranking
公开(公告)号：GB2579262B
公开(公告)日：2021-02-10
申请号：GB201911502
申请日：2019-08-12
Applicant: ADOBE INC
Inventor： JOON-YOUNG LEE , NING XU , SEOUNGWUG OH
IPC: G06T7/11
Abstract: A space-time memory (i.e. neural) network locates target object(s) in video content for segmentation or other object classification. Claimed is generating (204) a query key map and a query value map by applying a space-time memory network to a query frame (202) depicting a target feature and retrieving (206), from a memory, a memory key map and a memory value map that are computed from a set of memory frames from video content that includes the query frame. Memory weights are computed (208) by applying a similarity function to the memory key map and the query key map and the space-time memory network is used to classify (210) (e.g. generate a segmentation mask) content in the query frame as depicting the target feature based on a weighted summation that includes the memory weights applied to memory locations in the memory value map. Also claimed is accessing, from video content, a query frame having content depicting a target feature, and performing a step for classifying content of the query frame as depicting the target feature by applying a space-time memory (i.e. neural) network to the query frame and one or more memory frames.
-
公开(公告)号：GB2579262A
公开(公告)日：2020-06-17
申请号：GB201911502
申请日：2019-08-12
Applicant: ADOBE INC
Inventor： JOON-YOUNG LEE , NING XU , SEOUNGWUG OH
IPC: G06T7/11
Abstract: A space-time memory (i.e. neural) network locates target object(s) in video content for segmentation or other object classification. Claimed is generating (204) a query key map and a query value map by applying a space-time memory network to a query frame (202) depicting a target feature and retrieving (206), from a memory, a memory key map and a memory value map that are computed from a set of memory frames from video content that includes the query frame. Memory weights are computed (208) by applying a similarity function to the memory key map and the query key map and the space-time memory network is used to classify (210) (e.g. generate a segmentation mask) content in the query frame as depicting the target feature based on a weighted summation that includes the memory weights applied to memory locations in the memory value map. Also claimed is accessing, from video content, a query frame having content depicting a target feature, and performing a step for classifying content of the query frame as depicting the target feature by applying a space-time memory (i.e. neural) network to the query frame and one or more memory frames.
-
公开(公告)号：GB2582689B
公开(公告)日：2022-03-16
申请号：GB201918840
申请日：2019-12-19
Applicant: ADOBE INC
Inventor： BRIAN LYNN PRICE , SCOTT COHEN , MARCO FORTE , NING XU
Abstract: Techniques are disclosed for deep neural network (DNN) based interactive image matting. A methodology implementing the techniques according to an embodiment includes generating, by the DNN, an alpha matte associated with an image, based on user-specified foreground region locations in the image. The method further includes applying a first DNN subnetwork to the image, the first subnetwork trained to generate a binary mask based on the user input, the binary mask designating pixels of the image as background or foreground. The method further includes applying a second DNN subnetwork to the generated binary mask, the second subnetwork trained to generate a trimap based on the user input, the trimap designating pixels of the image as background, foreground, or uncertain status. The method further includes applying a third DNN subnetwork to the generated trimap, the third subnetwork trained to generate the alpha matte based on the user input.
-
公开(公告)号：GB2560219B
公开(公告)日：2021-01-27
申请号：GB201718547
申请日：2017-11-09
Applicant: ADOBE INC
Inventor： BRIAN LYNN PRICE , STEPHEN SCHILLER , SCOTT COHEN , NING XU
Abstract: Methods and systems are provided for generating mattes for input images. A neural network system can be trained where the training includes training a first neural network that generates mattes for input images where the input images are synthetic composite images. Such a neural network system can further be trained where the training includes training a second neural network that generates refined mattes from the mattes produced by the first neural network. Such a trained neural network system can be used to input an image and trimap pair for which the trained system will output a matte. Such a matte can be used to extract an object from the input image. Upon extracting the object, a user can manipulate the object, for example, to composite the object onto a new background.
-
公开(公告)号：GB2582689A
公开(公告)日：2020-09-30
申请号：GB201918840
申请日：2019-12-19
Applicant: ADOBE INC
Inventor： BRIAN LYNN PRICE , SCOTT COHEN , MARCO FORTE , NING XU
Abstract: A method for using a deep neural network (DNN) to perform image matting comprises receiving an input image 110 and user input 105 specifying a foreground location in the image, generation of a binary mask 130 based on the user-specified foreground location, generation of a trimap 150 based on the user-specified foreground location and the binary mask, and generation of an alpha matte 170 based on the trimap and the user-specified foreground location. The generation of the binary mask, trimap, and alpha matte are performed by the DNN 120, 140, 160. The training of the DNN may minimise a loss function based on a comparison of the generated alpha matte pixels for a training image and ground-truth alpha matte pixels associated with the image. The method may comprise down-sampling the input image to produce a second image and the production of a first alpha matte based on the original image and a second alpha matte based on the second image. The DNN may generate a foreground and/or background colour decontamination map providing colour channels associated with the alpha matte. A computer program for the production of trimaps using input images, binary masks, and user inputs is also disclosed.
-
-
-
-
-
-
-
Search Results
8
records
(took 0.02 seconds)
