摘要 |
<P>PROBLEM TO BE SOLVED: To more stably generate a three-dimensional image from an input image of a two-dimensional image. <P>SOLUTION: A scene classification processing section 21 detects each adaptation for a plurality of scenes to the input image consisting of the two-dimensional image. A time base control processing section 22 generates a mixture ratio normalized from each adaptation for the plurality of scenes, and supplies it to a blend processing section 23. A depth image generation section for every classified scene 24 generates a depth image from the input image by a method at every scene. The blend processing section 23 blends the depth images generated by the method at every scene at a ratio depending on the adaptation at every scene to generate a mixed depth image. A parallax generation processing section 25 generates the three-dimensional image which consists of a left eye image and a right eye image, which are obtained by shifting respective pixels of the input image, and can three-dimensionally be viewed by using the mixed depth image. This technique can be applied to a three-dimensional image processor. <P>COPYRIGHT: (C)2013,JPO&INPIT |