Smart cropping of video images in a videoconferencing session
AssigneePolycom, Inc.
InventorsNiv Wagner, Zigi Gavish
Life of this patent
Estimating the term…
Abstract
Systems and methods are disclosed for controlling cropping areas of video images to match the allocated area associated with the image in a video conferencing layout. The disclosed methods can protect regions of interest from being cropped. A region of interest within a video image is identified to adjust the cropping in such a way that the region of interest is preserved within the cropped image. The region of interest may be identified based on motion detection or flesh tone detection, for example.
Claims21
A method of formatting a video image associated with an endpoint for display in a videoconference, the method comprising:
- determining a region of interest within the video image;
- and
- cropping the video image to match an associated segment in a conferencing image, wherein the cropping preserves the region of interest.
The method of 1 wherein determining a region of interest comprises both motion and color analyzing.
An apparatus for conducting a videoconference, comprising:
- one or more region of interest detector modules configured to determine a region of interest within a video image;
- and
- a cropping manager module configured to define cropping instructions for video image to create a cropped image wherein the determined region of interest is within the cropped image.
An apparatus for conducting a videoconference, comprising:
- one or more input modules adapted to receive and decode compressed video images transmitted by a video stream;
- one or more region of interest detector modules configured to determine a region of interest within a decoded a video image;
- a cropping manager module configured to define cropping instructions for the decoded video image to create a cropped image wherein the determined region of interest is within the cropped image;
- and
- one or more output modules is adapted to generate a continuous presence video image from one or more of the decoded video images streams wherein the continuous presence video image comprises the cropped image.
Dependencies are inferred from the claim text.