This paper presents a new method of converting 2-D video into 3-D stereoscopic video based upon the motion type of key frames. The proposed conversion method provides the stable display of 3-D stereoscopic depth. To realize this, we propose the concep...
This paper presents a new method of converting 2-D video into 3-D stereoscopic video based upon the motion type of key frames. The proposed conversion method provides the stable display of 3-D stereoscopic depth. To realize this, we propose the concept of key frame that has the reliability of motion vector field(MVF) as well as the accuracy of estimated motion type. If a current frame does not belong to the key frame, its asscciated motion type follows the key frame. Therefore, the abrupt variation of the motion type can be avoided. As a result, the stereoscopic conversion can be smoothly varied between subsequent frames. Experimental results show that the stereoscopic conversion of any video can be shown in the stable manner.