SCENE SEPARATION AND SYNTHESIS PROCESSING IN RAY-SPACE(International Workshop on Advanced Image Technology 2007)

概要

論文の詳細を見る
This paper proposes a novel method to remove obstacles and synthesize scenes by using multi-view images. An arbitrary viewpoint image can be synthesized by using the Ray-Space method, so we can choose any image that we want to see. However, if the object that we want to see is concealed by an obstacle, we can't see it even if we use the Ray-Space method. So, by removing the obstacles, we can generate images with the object that we want to see. Additionally, by synthesizing images captured in different situations, we can generate an image with objects as if they exist in the same situation. In this research, we use the EPI which is the horizontal cross section of Ray-Space. On the EPI, the objects are represented by straight lines, and the inclination of the straight line depends on the depths of the objects. So, if straight lines of a specific inclination can be detected, it will be possible to detect the objects that exist in a specific depth. Therefore, if the straight lines that represent the obstacle are removed, the obstacle will also be removed. But it is necessary to bridge the part where the obstacle was removed by the color of the object behind the obstacle. And, if the straight lines that represent an object is cut out and put on another EPI, the object would be synthesized using other images. However, the added straight lines must not be placed over the straight lines that represent the object which is in front of the new object. Thus, the required image can be synthesized by combining these techniques with the Ray-Space method.
社団法人電子情報通信学会の論文
2007-01-03

SCENE SEPARATION AND SYNTHESIS PROCESSING IN RAY-SPACE(International Workshop on Advanced Image Technology 2007)

スポンサーリンク

概要

著者

関連論文

スポンサーリンク