US 12,445,726 B2
Stabilization of video with object detection
Séraphin Cyrille Philippe Bonnaffé, Paris (FR); and Nicolas Rahmouni, Paris (FR)
Assigned to GoPro, Inc., San Mateo, CA (US)
Filed by GoPro, Inc., San Mateo, CA (US)
Filed on Nov. 24, 2023, as Appl. No. 18/518,896.
Application 18/518,896 is a continuation of application No. 17/739,742, filed on May 9, 2022, granted, now 11,889,189.
Prior Publication US 2024/0089600 A1, Mar. 14, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. H04N 23/68 (2023.01); G02B 27/64 (2006.01); H04N 5/14 (2006.01)
CPC H04N 23/683 (2023.01) [G02B 27/646 (2013.01); H04N 5/144 (2013.01); H04N 23/6811 (2023.01)] 20 Claims
OG exemplary drawing
 
1. An image capture device for stabilizing videos, the image capture device comprising:
an image sensor configured to capture visual content during a capture duration based on light received within a field of view of the image sensor; and
one or more physical processors configured by machine-readable instructions to:
switch between movement stabilization and object stabilization to stabilize the visual content captured at different moments within the capture duration based on whether or not the visual content captured at the different moments within the capture duration depicts an object of interest, wherein the movement stabilization is used to stabilize the visual content captured at a first moment in the capture duration based on the visual content captured at the first moment not depicting the object of interest and the object stabilization is used to stabilize the visual content captured at a second moment in the capture duration based on the visual content captured at the second moment depicting the object of interest, wherein the switch between the movement stabilization and the object stabilization to stabilize the visual content of the video includes:
determination of whether the object of interest is depicted within the visual content captured at a given moment during the capture duration;
responsive to determination that the visual content captured at the given moment does not depict the object of interest, stabilization of the visual content captured at the given moment based on the movement stabilization, the movement stabilization including stabilization of the visual content based on movement of the image capture device, the movement stabilization of the visual content including use of a first viewing window to provide a first punchout of the visual content to stabilize the visual content, wherein the movement stabilization does not take into account depiction of things within the visual content and stabilizes the visual content without regards as to relationship between where the first viewing window is placed within the visual content and where things are depicted within the visual content; and
responsive to determination that the visual content captured at the given moment depicts the object of interest, stabilization of the visual content captured at the given moment based on the object stabilization, the object stabilization including use of placement of the object of interest within the visual content to determine placement of a second viewing window to provide a second punchout of the visual content that keeps the object of interest within a center portion of the stabilized visual content, wherein the object stabilization does not take into account movement of the image capture device, the second viewing window being smaller or having a different shape than the first viewing window.