Abstract
3D video makes it possible for audience to enjoy stereo perception by means of multi-view rendering. Revealing a more realistic and vivid world that conventional 2D video can never achieve, this innovation becomes an important direction of the development of multimedia in the future, thus converting existed 2D video into 3D ones has its great significance. This paper presents a semi-auto 2D-to-3D video conversion method, which segments a video flow into key frames and non-key frames, then utilizes an interactive algorithm as well as an efficient edge detection technique to acquire depth map for a single key frame. Depth map for non-key frame is derived by motion tracking and compensation automatically. The final 3D video flow can be rendered based on the depth map calculated above.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call