-
公开(公告)号：KR101267629B1
公开(公告)日：2013-05-27
申请号：KR1020110018563
申请日：2011-03-02
Applicant: 한국과학기술원
CPC classification number: G06T11/203
Abstract: 밀집된데이터를갖는 1차원그래프간소화방법, 장치, 시스템및 이를위한기록매체를공개한다. 본발명은원본그래프를인가받아다이나믹프로그래밍기법을이용하여원본그래프상의지정된영역에서연결가능한두 데이터점을이어주는복수개의직선들각각과원본그래프사이의가장큰 수직거리를나타내는코스트를최소화시켜주는점들인복수개의특징점이추출되는단계, 및추출된특징점이지정된방식으로연결되고기울기가조절되어간소화그래프가출력되는단계를구비한다. 따라서 1차원그래프를간소화함에있어서최적의결과를보장하며사용자의상호작용적제어를가능하게한다.
-
公开(公告)号：KR101261737B1
公开(公告)日：2013-05-09
申请号：KR1020110056907
申请日：2011-06-13
Applicant: 한국과학기술원
IPC: G06T13/00
Abstract: 얼굴애니메이션리타게팅방법및 이를위한기록매체를공개한다. 본발명은하나의타겟모델에대해 RBF 기반회기트레이닝및 kCCA 기반회기트레이닝을각각 1회씩수행하고, 이후 RBF 기반회기트레이닝결과및 kCCA 기반회기트레이닝결과를조합함에의해형태학적으로매우상이한타겟모델에대해서도소스애니메이션의표정에대응하는매우자연스러운표정을회득할 수있다. 또한모션와핑을추가함에의해타겟모델의개성을더욱두드러지게표현할수 있다.
-
公开(公告)号：KR1020130010408A
公开(公告)日：2013-01-28
申请号：KR1020110071192
申请日：2011-07-18
Applicant: 한국과학기술원
CPC classification number: H04N2013/0074
Abstract: PURPOSE: A disparity graph compression control device of a stereoscopic image, a method, an image processing device and a recording medium are provided to supply environment capable of disparity design corresponding to storytelling to an operator. CONSTITUTION: An initial disparity graph generation unit(120) generates an one-dimensional initial disparity graph for a stereoscopic image. A total compression unit(131) compresses the initial disparity graph with a Poisson equation. A scene connection unit(140) connects the compressed disparity graph by the total compression unit. [Reference numerals] (100) Input unit; (111) Left image buffer; (112) Right image buffer; (120) Initial disparity graph generation unit; (131) Total compression unit; (132) Partial correction unit; (140) Scene connection unit; (150) Rendering unit; (AA) Stereoscopic image
Abstract translation: 目的：提供立体图像的视差图形压缩控制装置，方法，图像处理装置和记录介质，以向操作者提供能够对应于讲故事的视差设计的环境。 构成：初始视差图生成单元（120）生成立体图像的一维初始视差图。 总压缩单元（131）用泊松方程压缩初始视差图。 场景连接单元（140）将压缩视差图连接到总压缩单元。 （附图标记）（100）输入单元; （111）左图像缓冲区; （112）右图像缓冲; （120）初始视差图生成单元; （131）总压缩机组; （132）部分校正单位; （140）场景连接单元; （150）渲染单元; （AA）立体图像
-
公开(公告)号：KR101181199B1
公开(公告)日：2012-09-18
申请号：KR1020110000260
申请日：2011-01-03
Applicant: 한국과학기술원
IPC: G06T17/05
CPC classification number: G06T17/05 , G06T17/205
Abstract: Disclosed herein are a stereoscopic image generation method of background terrain scenes, a system using the same, and a recording medium for the same. The stereoscopic image generation method of background terrain scenes includes an initial mesh creation step of creating an initial mesh using terrain geometry based on image sequences, a geometry error correction step of generating a projection map, detecting error regions of the initial mesh using the generated projection map, generating a vector map of the detected error regions, and generating a corrected mesh, error of which is corrected, and a stereo conversion step of generating a stereoscopic image using the corrected mesh. Since the stereoscopic image is generated based on the mesh, the mesh fits the terrain shape even though the geometry is complex. Further, time coherence can be enforced, the mesh can be edited easily, and new elements can be unseamingly composed into the terrain. Thus, it is possible to prevent a viewer who views the stereoscopic image from becoming tired.
-
公开(公告)号：KR1020120078949A
公开(公告)日：2012-07-11
申请号：KR1020110000260
申请日：2011-01-03
Applicant: 한국과학기술원
IPC: G06T17/05
CPC classification number: G06T17/05 , G06T17/205
Abstract: PURPOSE: A method for generating stereoscopic images about background terrain scenes, a system therefor, and a recording medium therefor are provided to generate stereoscopic images based on a mesh and fix a topography shape even though geometry is complicated. CONSTITUTION: A system generates an initial mesh for generating terrain geometry using a radial basic function technique(S100). The system corrects an error of the initial mesh using a projection map(S200). The system generates a depth map based on the corrected mesh. The system adjusts the depth of a combined target object(S300).
Abstract translation: 目的：提供一种用于产生关于背景地形场景的立体图像，其系统及其记录介质的方法，用于基于网格生成立体图像，并且即使几何形状复杂也能固定形状形状。 构成：系统使用径向基本功能技术生成用于生成地形几何的初始网格（S100）。 系统使用投影图来校正初始网格的错误（S200）。 系统基于校正的网格生成深度图。 系统调整组合目标对象的深度（S300）。
-
Patent Agency Ranking
公开(公告)号：KR1020110059275A
公开(公告)日：2011-06-02
申请号：KR1020090115949
申请日：2009-11-27
Applicant: 한국과학기술원
IPC: G06T15/50
CPC classification number: G06T15/506 , G06K9/6218 , G06T13/40 , G06T15/60 , G06T2207/10152 , G06T2215/12
Abstract: PURPOSE: A system for editing an interactive lighting effect for a 3D rendered image is provided to automatically determining the optimal number of lightings and parameters of each lighting device. CONSTITUTION: A sketching interface(71) inputs an illuminating effect to a vertex of a 3D object. A parameter control interface(72) controls all kinds of parameters about the mapping data of the 3D object. A data clustering unit(74) calculates the optimal number of clusters by the clustering process. A lighting estimation unit(75) presumes the parameters of the spot lightings. A rendering unit(76) creates a scene in which a illuminating effect is applied to the 3D object.
Abstract translation: 目的：提供用于编辑3D渲染图像的交互式照明效果的系统，以自动确定每个照明设备的最佳照明和参数数量。 构成：素描界面（71）向3D对象的顶点输入照明效果。 参数控制接口（72）控制关于3D对象的映射数据的各种参数。 数据聚类单元（74）通过聚类处理计算聚类的最佳数量。 照明估计单元（75）假设照明灯的参数。 渲染单元（76）创建将照明效果应用于3D对象的场景。
-
-
公开(公告)号：KR101898512B1
公开(公告)日：2018-09-14
申请号：KR1020170079893
申请日：2017-06-23
Applicant: 한국과학기술원
Abstract: 본발명의실시예에따른모션변환장치는, 기본모션데이터가입력되는기본데이터입력부; 입력된기본모션데이터의전처리를통해파라미터를획득하는전처리부; 상기전처리된파라미터와연령입력정보에기초하여, 생체역학적관찰기반의보행동작변환처리를수행하는모션변환부; 및상기변환처리된모션을출력하는출력부를포함한다.
-
公开(公告)号：KR1020170133836A
公开(公告)日：2017-12-06
申请号：KR1020160065131
申请日：2016-05-26
Applicant: 한국과학기술원
IPC: B25J9/16
Abstract: 모델예측제어방법이개시된다. 모델예측제어방법은, 캐릭터에대한시스템동역학모델을입력받는단계, 미분정보의보간기법을이용하여, 입력된시스템동역학모델에대한수치적미분을수행하는단계및 시스템동역학모델에대한미분정보를출력하는단계를포함한다.
Abstract translation: 公开了一种模型预测控制方法。 模型预测控制方法包括以下步骤：接收字符的系统动力学模型;使用差分信息的插值技术对输入系统动力学模型执行数值微分;以及输出关于系统动力学模型的差分信息 和的步骤。
-
公开(公告)号：KR1020170098135A
公开(公告)日：2017-08-29
申请号：KR1020160054159
申请日：2016-05-02
Applicant: 한국과학기술원
CPC classification number: G06T19/20 , G06T15/08 , H04N13/275
Abstract: 본발명은깊이이미지를이용한다시점객체분리방법및 시스템에관한것으로, 이는기준컬러이미지와깊이이미지를시점일치시킨후, 상기깊이이미지를기반으로타겟이미지각각에대응되는깊이이미지들을획득하고, 이들을기반으로시점간대응영역을생성하는시스템셋업단계; 및상기타겟이미지각각의컬러정보를기반으로단일이미지에너지항을생성하고, 상기시점간대응영역을기반으로시점간컬러와깊이정보차이를반영하는시점간 에너지항을추가로생성한후, MRF(Markov Random Field) 에너지최적화를통해상기타겟이미지각각에포함된객체를분리하는객체분리단계를포함하여, 시점간일관성을유지하여객체분리동작을수행할수 있도록한다.
Abstract translation: 本发明涉及一种点对象分离方法和系统，用于深度使用图像，其被配置为获得参考颜色的图像，然后匹配的深度图像，并且对应于目标图像的深度图像时，分别基于深度图像上，那些基于 系统设置步骤，生成视点之间的对应区域; 并且基于每个目标图像的颜色信息生成单个图像能量项，并且进一步基于视图间对应区域生成反映视图间颜色和深度信息之间的差异的视图间能量项， 以及对象分离步骤，通过马尔可夫随机场的能量优化来分离包括在每个目标图像中的对象，从而可以在维持视点之间的一致性的同时执行对象分离操作。
-
-
-
-
-
-
-
-
-
Search Results
98
records
(took 0.026 seconds)
