Invention Grant
- Patent Title: Segmentation-guided real-time facial performance capture
-
Application No.: US15438551Application Date: 2017-02-21
-
Publication No.: US10198624B2Publication Date: 2019-02-05
- Inventor: Hao Li , Tianye Li , Shunsuke Saito
- Applicant: Pinscreen, Inc.
- Applicant Address: US CA Santa Monica
- Assignee: Pinscreen, Inc.
- Current Assignee: Pinscreen, Inc.
- Current Assignee Address: US CA Santa Monica
- Agency: Socal IP Law Group LLP
- Agent Jonathan Pearce
- Main IPC: G06K9/00
- IPC: G06K9/00 ; G06T7/11 ; G06K9/46 ; G06T7/143 ; G06T7/194

Abstract:
There is disclosed a system and method of performing facial recognition from RGB image data. The method includes generating a lower-resolution image from the RGB image data, performing a convolution of the lower-resolution image data to derive a probability map identifying probable facial regions and a probable non-facial regions, and performing a first deconvolution on the lower-resolution image using a bilinear interpolation layer to derive a set of coarse facial segments. The method further includes performing a second deconvolution on the lower-resolution image using a series of unpooling, deconvolution, and rectification layers to derive a set of fine facial segments, concatenating the set of coarse facial segments to the set of fine facial segments to create an image matrix made up of a set of facial segments, and generating a binary facial mask identifying probable facial regions and probable non-facial regions from the image matrix.
Public/Granted literature
- US20170243053A1 REAL-TIME FACIAL SEGMENTATION AND PERFORMANCE CAPTURE FROM RGB INPUT Public/Granted day:2017-08-24
Information query