公开(公告)号：US10511846B1

公开(公告)日：2019-12-17

申请号：US15254723

申请日：2016-09-01

Applicant: GOOGLE INC.

Inventor： Jing Chen ,  Marco Paniconi ,  Michael Horowitz

IPC: H04N19/117 ,  H04N19/137 ,  H04N19/36 ,  H04L12/801 ,  H04N19/14 ,  H04N19/176 ,  H04N19/172 ,  H04N19/182

Abstract: A method and apparatus for adaptive denoising of source video in a video conference application is provided. Source video frames received from a video capture device are divided into a plurality of 16×16 blocks. For each source block, a moving object detection process and a noise estimation process are performed. Then temporal denoising is adaptively applied to the blocks of the source frame based on the noise estimation and moving object detection. The adaptively filtered blocks are provided to an output frame and forwarded to a coding module for encoding.

公开(公告)号：US10049436B1

公开(公告)日：2018-08-14

申请号：US15232212

申请日：2016-08-09

Applicant: GOOGLE INC.

Inventor： Jing Chen ,  Marco Paniconi

IPC: H04N7/15 ,  G06T5/00 ,  H04N7/14

Abstract: A method and apparatus for adaptive denoising of source video in a video conference application is provided. Video captured is analyzed on a frame by frame basis to determine whether denoising of the frame should be performed prior to providing the source frame to an encoder. If the frame is to be denoised, the frame is divided into a plurality of blocks and a local denoising process is performed on a block per block basis.