Abstract:
The present invention relates to a depth information extracting method and a stereoscopic image editing method using the same method comprising the following steps of: extracting proxy data from left/right images; generating an edit decision list after drafting a sequence by selecting images based on the extracted proxy data; computing binocular disparity data of left and light images of the drafted sequence; generating a disparity map using the computed binocular disparity data; generating a depth map using the disparity map; and finally editing the drafted sequence based on the disparity map, the depth map and the edit decision list. The present invention allows a user to edit more objective depth, thereby improving the completeness of an image and solving a problem about a human factor. [Reference numerals] (S110) Photograph right/left images;(S120) Extract a proxy;(S130) Draft a sequence;(S140) Extract depth information;(S142) Draft depth information;(S144) Indicate 2D-3D;(S150) Edit the sequence;(S160) D.I & mastering
Abstract:
본 발명은 컬러세그먼트 추출과 깊이지도를 활용한 입체영상 제작 방법에 관한 것으로서, (a) 단안카메라로 촬영된 소스영상에서 깊이지도의 샘플링을 통해 분석대상이 되는 프레임범위를 설정하는 단계;와 (b) 상기 소스영상 속에 보여지는 모든 개체들을 분석하여 전경과 원경 및 배경을 포함하는 요소(element)의 공간구조를 분석하고, 상기 분석된 요소에 컬러속성을 부여하는 단계; (c) 상기 소스영상에서 모션벡터를 산출하여 상기 요소의 움직임을 분석한 후에, 상기 소스영상을 상기 분석된 요소별로 분리된 컬러세그먼트영상으로 변환하는 단계; (d) 상기 컬러세그먼트영상을 그레이스케일로 된 깊이지도(depth map)로 변환하는 단계; (e) 상기 변환된 깊이지도를 이용하여 상기 소스영상을 양안영상으로 제작하는 단계; 및 (f) 상기 (d)단계에서의 깊이지도의 깊이값을 보정한 후에 상기 (e)단계를 재수행하는 단계;를 포함한다.
Abstract:
PURPOSE: A stereoscopic image production method using a color segment extraction and a depth map is provided to produce a stereoscopic image more rapidly and easily by converting an image source photographed by the usage of a single-lens production technique into a stereoscopic image. CONSTITUTION: An image analysis depth script is formed by the analysis of each entity in a source image. A motion vector is calculated by the analysis and separation of each entity in a formed image. Based on the calculated motion vector, a depth map about foreground and background is generated in the source image. Based on the depth map, a three-dimensional image is generated. [Reference numerals] (AA) Forming an image analysis depth scrip; (BB) Calculating a motion vector; (CC) Generating a background depth map; (DD) Generating a foreground depth map; (EE) Forming a L.R stereo image; (FF) Correcting the depth map; (GG) Outputting a master