REALTIME PHOTO RETOUCHING OF LIVE VIDEO

摘要

Provided are techniques for processing a subset of a plurality of video frames to produce a 3D model of a particular element within each frame of the subset; processing the plurality of video frames to identify a particular feature within the particular element in each frame of the plurality of frames in which the particular feature appears to produce a plurality of instances of the element, each instance of the plurality of instances associated with a corresponding frame of the plurality of frames; modifying each instance of the plurality of instances in accordance with use parameters and in conformity with the 3D model to generate a plurality of modified instances; and replacing each instance of the plurality of instances with a corresponding modified instance in the corresponding frame of the plurality of frames to produce a modified video file.

主权项

1. A method, comprising:
processing a subset of a plurality of video frames to produce a 3D model of a particular element within each frame of the subset; processing the plurality of video frames to identify a particular feature within the particular element in each frame of the plurality of frames in which the particular feature appears to produce a plurality of instances of the element, each instance of the plurality of instances associated with a corresponding frame of the plurality of frames; modifying each instance of the plurality of instances in accordance with user-defined parameters and in conformity with the 3D model to generate a plurality of modified instances; and replacing each instance of the plurality of instances With a corresponding modified instance in the corresponding frame of the plurality of frames to produce a modified video file,