-
公开(公告)号：KR101355894B1
公开(公告)日：2014-01-29
申请号：KR1020120102279
申请日：2012-09-14
Applicant: 한밭대학교 산학협력단
IPC: H04N13/00
CPC classification number: H04N13/271 , H04N13/122
Abstract: The present invention relates to an apparatus and a method for converting a two-dimensional image into a three-dimensional image using adaptive depth fusion, which is capable of integrating various conversion clues existing in a two-dimensional image to perform scene-adaptive depth fusion, thereby effectively representing a three-dimensional image of a current scene. The apparatus according to the present invention comprises: a video source input unit for providing a two-dimensional image; a global depth generator for generating global depth in the two-dimensional image based on geometric information; an object depth generator for generating the object depth of an object area in the two-dimensional image based on focus information or motion information; and a depth fusion unit for fusing the global depth and the object depth, wherein the depth fusion unit fuses, if the object depth generator generates object depth based on the focus information, the global depth and the object depth by a weighted sum method and, if the object depth generator generates object depth based on motion information, fuses the global depth and the object depth by an independent depth allocation method. [Reference numerals] (10) Video source input unit; (20) Global depth generator; (30) Object depth generator; (40) Depth fusion unit
Abstract translation: 本发明涉及一种使用自适应深度融合将二维图像转换为三维图像的装置和方法，其能够集成存在于二维图像中的各种转换线索以执行场景自适应深度融合 从而有效地表示当前场景的三维图像。 根据本发明的装置包括：用于提供二维图像的视频源输入单元; 全球深度生成器，用于基于几何信息在二维图像中产生全局深度; 对象深度生成器，用于基于焦点信息或运动信息生成二维图像中的对象区域的对象深度; 以及深度融合单元，用于融合全局深度和物体深度，其中如果物体深度生成器通过加权和方法基于聚焦信息，全局深度和物体深度生成对象深度，则深度融合单元融合， 如果对象深度生成器基于运动信息生成对象深度，则通过独立的深度分配方法来融合全局深度和对象深度。 （附图标记）（10）视频源输入单元; （20）全球深度发生器; （30）物体深度发生器; （40）深度融合单元
-
公开(公告)号：KR101336577B1
公开(公告)日：2013-12-03
申请号：KR1020120064996
申请日：2012-06-18
Applicant: 한국항공대학교산학협력단
IPC: H04N19/50
CPC classification number: H04N19/513 , H04N19/176
Abstract: The present invention relates to an apparatus for inducing a motion vector and an induction method for the same, which is capable of inducing an additional MVP in a left candidate unit or inducing an MVP from an induction candidate group if an upper candidate unit MVP of a prediction unit is not valid when predicting two or more spatial MVPs. According to the present invention, the induction method using the apparatus comprising a motion vector induction unit, a redundancy check unit, a size limit unit and a motion vector addition unit comprises the steps of: (c) combining a left candidate block and an upper candidate block adjacent to a prediction unit to be encoded or decoded to form an induction candidate group when inducing a spatial MVP using the motion vector induction unit; (d) setting the induction conditions of the induction candidate group using the motion vector induction unit; and (e) determining the induction conditions to induce an MVP for each induction candidate group, wherein the (c) step includes a scaling condition setting step for determining whether to scale the induction candidate group. [Reference numerals] (AA) Step S200
Abstract translation: 本发明涉及一种用于诱导运动矢量的装置及其感应方法，其能够在左候选单元中引入附加的MVP或者从感应候选组中诱导MVP，如果上级候选单元MVP为 预测单位在预测两个或多个空间MVP时无效。 根据本发明，使用包括运动矢量感应单元，冗余校验单元，大小限制单元和运动矢量相加单元的装置的感应方法包括以下步骤：（c）组合左候选块和上位 候选块与预测单元相邻以被编码或解码以在使用运动矢量感应单元诱导空间MVP时形成感应候选组; （d）使用运动矢量感应单元设定感应候选组的诱导条件; （e）确定诱导条件以诱导每个诱导候选组的MVP，其中（c）步骤包括用于确定是否缩放感应候选组的缩放条件设置步骤。 （标号）（AA）步骤S200
-
公开(公告)号：KR101237411B1
公开(公告)日：2013-02-27
申请号：KR1020110031265
申请日：2011-04-05
Applicant: 한국항공대학교산학협력단
IPC: H04N19/70 , H04N19/103
Abstract: 본 발명은 비디오 압축 표준인 H.264/AVC와 비디오부호화연합팀(JCT-VC)에서 표준화 중인 차세대 동영상 압축기술(HEVC)의 신택스 구조에 관한 것으로, 본 발명에 따른 비디오 코덱의 예측단위 신택스 시스템을 이용한 예측단위 구분 방법은 분할여부통지부를 이용하여, 신택스 요소에 부호 없는 정수 1비트를 할당하여, 현재 예측단위의 분할여부를 통지하는 단계, 예측모드통지부를 이용하여, 분할 여부에 따라 신택스 요소에 부호 없는 정수 2비트 또는 정수1비트를 할당하여, 예측 모드를 통지하는 단계 및 예측단위통지부를 이용하여, 예측단위별 확률 분포를 조사하여, 부호 없는 정수 2비트 또는 부호 없는 정수 지수 골룸 코드를 할당하여, 예측단위를 통지하는 단계를 포함하는 기술적 것을 특징으로 한다.
-
公开(公告)号：KR1020120113518A
公开(公告)日：2012-10-15
申请号：KR1020110031265
申请日：2011-04-05
Applicant: 한국항공대학교산학협력단
IPC: H04N19/70 , H04N19/103
CPC classification number: H04N19/70 , H04N19/103 , H04N19/184
Abstract: PURPOSE: A prediction unit syntax system of a video codec and a prediction unit classifying method using the same are provided to supply a syntax structure which can efficiently express the size of a block of a prediction unit. CONSTITUTION: A division determination notifying unit(10) notifies division determination of a current prediction unit by parsing a prediction unit bit stream. A prediction mode notifying unit(20) is connected to the division determination notifying unit. The division determination notifying unit notifies a division mode of the prediction unit. A prediction unit notifying unit(30) is connected to the prediction mode notifying unit. The prediction unit notifying unit notifies a prediction unit of a prediction mode. [Reference numerals] (10) Division determination notifying unit; (20) Prediction mode notifying unit; (30) Prediction unit notifying unit
Abstract translation: 目的：提供视频编解码器的预测单元语法系统和使用其的预测单元分类方法，以提供可以有效地表达预测单元的块的大小的语法结构。 构成：分割确定通知单元（10）通过解析预测单元比特流来通知当前预测单元的分割确定。 预测模式通知单元（20）连接到分割确定通知单元。 分割确定通知单元通知预测单元的分割模式。 预测单元通知单元（30）连接到预测模式通知单元。 预测单元通知单元向预测单元通知预测模式。 （附图标记）（10）分割判定通知单元; （20）预测模式通知单元; （30）预测单元通知单元
-
公开(公告)号：KR101336578B1
公开(公告)日：2013-12-03
申请号：KR1020120062054
申请日：2012-06-11
Applicant: 한국항공대학교산학협력단
IPC: H04N19/59
CPC classification number: H04N19/593 , H04N19/11 , H04N19/159 , H04N19/176
Abstract: The present invention relates to a most probable mode (MPM) inducing method which is used for encoding/decoding a prediction mode in a screen in a video encoding method. The prediction method of a prediction mode in a screen according to the present invention comprises the steps of (a) inducing an MPM using a most probable mode inducing unit and (b) encoding or decoding the induced MPM or a prediction unit in the screen of a current prediction unit using encoding/decoding units, wherein when two or more MPMs are induced from a neighboring prediction unit, two or more MPMs on a certain prediction unit in an LCU are induced in the (a) step, and wherein the prediction units except for the certain prediction unit in the LCU share two or more MPMs on the certain prediction unit. [Reference numerals] (210) Same phrase encoding/decoding unit;(220) Consistent phrase encoding/decoding unit;(230) Realignment unit;(240) Self-encoding/decoding unit
Abstract translation: 本发明涉及用于在视频编码方法的屏幕中对预测模式进行编码/解码的最可能模式（MPM）诱导方法。 根据本发明的屏幕中的预测模式的预测方法包括以下步骤：（a）使用最可能的模式诱导单元诱导MPM，以及（b）编码或解码在屏幕中的感应MPM或预测单元 使用编码/解码单元的当前预测单元，其中当从相邻预测单元感应出两个或更多个MPM时，在（a）步骤中诱导LCU中的某个预测单元上的两个或更多个MPM，并且其中预测单元 除了LCU中的某个预测单元在某个预测单元上共享两个或更多个MPM。 （210）相同短语编码/解码单元;（220）一致的短语编码/解码单元;（230）重新对准单元;（240）自编码/解码单元
-
Patent Agency Ranking
公开(公告)号：KR1020210020984A
公开(公告)日：2021-02-24
申请号：KR1020210021568
申请日：2021-02-18
Applicant: 한국전자통신연구원
IPC: H04N19/50 , H04N19/59 , H04N19/107 , H04N19/44
Abstract: 영상부호화방법및 장치가제공된다. 영상의화면간 예측을수행을위하여컴퓨터에의해실행되는프로그램을기록하는기록매체에있어서, 상기프로그램은, 참조블록식별자를기반으로, 이전에복원된하나의복원블록을복호화대상블록의참조블록으로결정하는단계, 상기참조블록의움직임정보를기반으로상기복호화대상블록에적용되는움직임정보를결정하는단계, 및상기결정된움직임정보를기반으로상기복호화대상블록에대한복호화를수행하는단계를포함하되, 상기참조블록식별자는상기참조블록으로사용될복원블록을지시할수 있다.
-
-
-
公开(公告)号：KR101691901B1
公开(公告)日：2017-01-04
申请号：KR1020130078944
申请日：2013-07-05
Applicant: 한국전자통신연구원
CPC classification number: H04N19/105 , H04N19/11 , H04N19/157 , H04N19/176 , H04N19/46 , H04N19/503 , H04N19/593
Abstract: 영상부호화방법및 장치가제공된다. 본발명에따른화면내 예측을수행하는영상부호화방법은이전에복원된하나의참조블록이갖는인트라예측모드를기반으로부호화대상블록의인트라예측모드를결정하는단계, 상기결정된인트라예측모드를기반으로상기부호화대상블록에대한부호화를수행하는단계를포함하되, 상기참조블록은, 현재영상내에서상기부호화대상블록에가장근접한복원된블록들중 하나인것을특징으로한다.
Abstract translation: 提供了一种图像编码方法和装置。 当对片内的块进行图像编码时，片的恢复块中的至少一个块被设置为参考块。 当这样做时，参考块的编码参数被区分，并且要编码的块基于编码参数自适应地编码。
-
公开(公告)号：KR101678409B1
公开(公告)日：2016-11-22
申请号：KR1020160010152
申请日：2016-01-27
Applicant: 한국전자통신연구원
IPC: H04N19/105 , H04N19/109 , H04N19/174 , H04N19/176
CPC classification number: H04N19/105 , H04N19/11 , H04N19/157 , H04N19/176 , H04N19/46 , H04N19/503 , H04N19/593
Abstract: 영상부호화방법및 장치가제공된다. 영상의화면간 예측을수행을위하여컴퓨터에의해실행되는프로그램을기록하는기록매체에있어서, 상기프로그램은, 참조블록식별자를기반으로, 이전에복원된하나의복원블록을복호화대상블록의참조블록으로결정하는단계, 상기참조블록의움직임정보를기반으로상기복호화대상블록에적용되는움직임정보를결정하는단계, 및상기결정된움직임정보를기반으로상기복호화대상블록에대한복호화를수행하는단계를포함하되, 상기참조블록식별자는상기참조블록으로사용될복원블록을지시할수 있다.
-
-
-
-
-
-
-
-
-
Search Results
252
records
(took 0.031 seconds)
