发明名称 |
Image object segmentation using examples |
摘要 |
Systems and methods are disclosed herein for using one or more computing devices to automatically segment an object in an image by referencing a dataset of already-segmented images. The technique generally involves identifying a patch of an already-segmented image in the dataset based on the patch of the already-segmented image being similar to an area of the image including a patch of the image. The technique further involves identifying a mask of the patch of the already-segmented image, the mask representing a segmentation in the already-segmented image. The technique also involves segmenting the object in the image based on at least a portion of the mask of the patch of the already-segmented image. |
申请公布号 |
US9607391(B2) |
申请公布日期 |
2017.03.28 |
申请号 |
US201514817731 |
申请日期 |
2015.08.04 |
申请人 |
Adobe Systems Incorporated |
发明人 |
Price Brian;Lin Zhe;Cohen Scott;Yang Jimei |
分类号 |
G06K9/34;G06T7/00;G06K9/62 |
主分类号 |
G06K9/34 |
代理机构 |
Kilpatrick Townsend & Stockton LLP |
代理人 |
Kilpatrick Townsend & Stockton LLP |
主权项 |
1. A method for automatically segmenting an object in an input image by referencing a dataset of already-segmented example images, the method comprising:
identifying, by one or more computing devices, multiple already-segmented example images in the dataset based on the multiple already-segmented example images being similar to the input image; identifying, by the one or more computing devices, a respective patch in each of the multiple already-segmented example images based on the respective patch being similar to an area of the input image including a patch of the input image; comparing, by the one or more computing devices, the input image with each respective patch of the multiple already-segmented example images to identify which respective patch of the multiple already-segmented example images is most similar to the area of the input image including the patch of the input image; identifying, by the one or more computing devices, a mask of the most similar patch of the already-segmented example images, the mask representing a segmentation in the corresponding already-segmented example image; and segmenting, by the one or more computing devices, the object in the input image based on at least a portion of the mask of the most similar patch of the corresponding already-segmented example image. |
地址 |
San Jose CA US |