Method and system for creating a 3D video from a monoscopic 2D video and corresponding depth information

摘要

Image sensors and a depth sensor of a monoscopic video sensing device are utilized to capture a 2D video and corresponding depth information. Regions of interest (ROIs) for the captured 2D video are selected based on the captured corresponding depth information. The monoscopic video sensing device selectively processes the captured 2D video and the captured corresponding depth information based on the selected ROIs. A 3D video is composed from the processed 2D video for display. The captured depth information that is synchronized to the captured 2D video is stored as metadata, and may be interpolated to match video resolution of the captured 2D video. The captured 2D video and the captured corresponding depth information are enhanced through scalable video coding. With 3D video rendering, relevant image/video components of the captured 2D video are selected based on the selected ROIs to compose the 3D video together with the corresponding depth information.

主权项

1. A method, comprising:
capturing a video via one or more image sensors of a monoscopic video sensing device; capturing corresponding depth information for said captured video via a depth sensor of said monoscopic video sensing device; synchronizing said captured corresponding depth information to said captured video; selecting one or more regions of interest in said captured video; and composing a modified video from said captured video and said captured corresponding depth information, wherein said composing comprises scalable video coding, which compresses said captured video and said captured depth information into a base layer video and an enhancement layer video respectively wherein, depth information corresponding to said one or more regions of interest is allocated more bits for compression than rest of said depth information.