ADAPTIVE VIDEO QUALITY FOR LARGE-SCALE VIDEO CONFERENCING
Media, methods, and systems are disclosed for adaptively adjusting video quality in a virtual event hosted by a virtual event hosting platform. A plurality of presenter video streams may be received at a video server. A change of state associated with a presenter video stream of the plurality of presenter video streams may be detected. In response to detecting the change of state, a change to a video display parameter for the presenter video stream may be requested by the video server. The video server may receive the updated presenter video stream from a presenter computing device associated with the presenter video stream. The video server may send updated presenter video stream to a plurality of presenter computing devices associated with the plurality of presenter video streams.
1 . A computer-implemented method for adaptively adjusting video quality in a virtual event, the method comprising:
receiving a video stream that is associated with a presenter computing device and that is encoded using a first codec;
detecting a motion associated with a presenter in the video stream using a machine learning algorithm;
in response to detecting the motion associated with the presenter, changing from the first codec to a second codec that is a higher quality codec than the first codec;
receiving an updated video stream from the presenter computing device that is encoded using the second codec; and
selectively forwarding at least a portion of the updated video stream to other computing devices.
2 . The method of claim 1 , wherein in response to detecting the motion of the presenter further changing a frame rate, bitrate, or a resolution of the video stream.
3 . The method of claim 1 , wherein detecting the motion of the presenter in the video stream comprises detecting lips, face, shoulders or hands motions.
4 . The method of claim 1 , wherein in response to detecting the motion of the presenter further applying a video-super resolution algorithm to the video stream.
5 . The method of claim 1 , wherein changing from the first codec to the second codec comprises changing to H.264.
6 . The method of claim 1 , wherein in response to detecting the motion of the presenter further cropping the video stream.
7 . The method of claim 1 , wherein in response to detecting the motion of the presenter further applying a Gaussian blur to a background.
8 . A computer-readable media storing computer-executable instructions that, when executed by a processor, cause the processor to perform a method for adaptively adjusting video quality in a virtual event, the method comprising:
receiving a video stream associated with a presenter computing device;
determining regions of interest amongst a plurality of regions in the video stream using a machine learning algorithm;
in response to detecting the regions of interest in the video stream, increasing a video quality for the regions of interest that differentiates from a video quality for other regions of the plurality of regions not in the regions of interest;
receiving an updated video stream from the presenter computing device with an increased video quality for the regions of interest; and
selectively forwarding at least a portion of the updated video stream to other computing devices.
9 . The computer-readable media of claim 8 , wherein determining the regions of interest in the video stream comprises determining a presenter's face.
10 . The computer-readable media of claim 9 , wherein in response to determining the presenter's face further encoding the presenter's face at a higher frame rate than a background.
11 . The computer-readable media of claim 8 , wherein in response to determining the regions of interest further applying a video super-resolution algorithm to the video stream.
12 . The computer-readable media of claim 8 , wherein in response to determining the regions of interest further changing the video stream to a higher quality codec.
13 . The computer-readable media of claim 8 , wherein determining the regions of interest in the video stream comprises detecting lips, face, shoulders or hands motions.
14 . The computer-readable media of claim 8 , wherein in response to determining the regions of interest further cropping a frame of the video stream.
15 . The computer-readable media of claim 9 , wherein in response to determining the regions of interest further changing a frame rate, bitrate, and/or a resolution of the video stream.
16 . A system for adaptively adjusting video quality, comprising:
a processor; and
a memory storing instructions that, when executed by the processor, cause the processor to perform the instructions to:
receiving a video stream associated with a presenter computing device;
detecting a motion in a background of the video stream;
in response to detecting the motion in the background of the video stream, cropping a frame of the video stream, wherein cropping the frame comprises reducing an encoded size of an updated video stream relative to the video stream by encoding the updated video stream with the frame without the background after said cropping;
receiving the updated video stream from the presenter computing device; and
selectively forwarding at least a portion of the updated video stream to other computing devices.
17 . The system of claim 16 , wherein in response to detecting the motion in the background further applying a Gaussian blur to the background.
18 . The system of claim 16 , wherein in response to detecting the motion in the background further switching to a static background.
19 . The system of claim 16 , wherein in response to detecting the motion in the background further switching to a static background.
20 . The system of claim 16 , wherein in response to detecting the motion in the background further freezing the background.