专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

61. 发明授权

KR101461496B1 예측 모드를 이용한 영상 부호화 방법 및 컴퓨터 판독 가능한 기록매체 有权
标题翻译：图像编码方法和计算机可读可重复使用方法的预测
公开(公告)号：KR101461496B1
公开(公告)日：2014-11-17
申请号：KR1020130075086
申请日：2013-06-28
申请人： 한국전자통신연구원 , 광운대학교 산학협력단 , 경희대학교 산학협력단
发明人： 최해철 , 정세윤 , 임성창 , 최진수 , 홍진우 , 심동규 , 오승준 , 안창범 , 박광훈 , 국승용 , 박시내 , 정광수
IPC分类号： H04N19/50
摘要： 본 발명은 영상 부호화 및 복호화 기술에 관한 것으로, 더욱 상세하게는 단방향 예측을 이용하여 영상을 부호화하거나 복호화하는 기술에 관한 것이다.
본 발명은 매크로 블록에 대한 예측 모드를 결정하는 단계, 상기 매크로 블록을 복수의 서브 블록으로 분할하는 단계, 상기 예측 모드에 따라서 상기 각 서브 블록을 부호화하는 단계를 포함하는 것을 특징으로 하는 영상 부호화 방법을 제공한다.

62. 发明公开

KR1020140082932A 다수 코어를 이용한 인터 예측 방법 및 장치 审中-实审
标题翻译：使用多核心的内部预测的方法和装置
公开(公告)号：KR1020140082932A
公开(公告)日：2014-07-03
申请号：KR1020130159430
申请日：2013-12-19
申请人： 광운대학교 산학협력단
发明人： 유성은 , 황태진 , 안용조 , 심동규
IPC分类号： H04N19/51 , H04N19/43
CPC分类号： H04N19/436 , H04N19/43 , H04N19/51
摘要： Disclosed are a method and an apparatus for performing inter prediction using multi-cores. The apparatus for performing inter prediction using the multi-cores includes: a motion vector extraction unit for predicting a motion vector for a present block by referring to a motion vector for the reference block in a reference frame corresponding to the present block in the present frame; a frame unit motion prediction unit for allocating at least one thread to each of blocks forming the present frame to perform motion prediction in the unit of a frame; and a motion vector expression unit for expressing the motion vector for the present block using the predicted motion vector and a differential motion vector. Accordingly, inter prediction operations are performed in parallel by using the multi-cores, so that encoding and decoding rates of videos can be increased.
摘要翻译：公开了一种使用多核执行帧间预测的方法和装置。用于使用多核执行帧间预测的装置包括：运动矢量提取单元，用于通过参考与当前帧中的当前块对应的参考帧中的参考块的运动矢量来预测当前块的运动矢量 ; 帧单元运动预测单元，用于将至少一个线程分配给形成当前帧的每个块，以帧为单位执行运动预测; 以及运动矢量表达单元，用于使用预测运动矢量和差分运动矢量来表示当前块的运动矢量。因此，通过使用多核并行执行帧间预测操作，从而可以增加视频的编码和解码速率。

63. 发明公开

KR1020140082915A 다중 계층 비디오의 계층 간 부호화/복호화 장치 및 방법 审中-实审
标题翻译：用于层间编码/解码的可分级视频的设备和方法
公开(公告)号：KR1020140082915A
公开(公告)日：2014-07-03
申请号：KR1020130103550
申请日：2013-08-30
申请人： 광운대학교 산학협력단
发明人： 장형문 , 김경혜 , 심동규
IPC分类号： H04N19/50 , H04N19/00
CPC分类号： H04N19/105 , H04N19/33 , H04N19/567
摘要： Disclosed are a device and a method for predicting an interlayer structure for high speed encoding of a scalable video. The method for predicting an interlayer structure for a multi-layered video according to the present invention includes the steps of: determining a reference block to compart a frame of a reference layer of a location corresponding to an improvement block to compart a frame of an improvement layer; and determining a limit depth of the improvement block with reference to structure information in which the determined reference block is divided. A device for interlayer encoding/decoding for a multi-layered video according to the present invention comprises: a corresponding block determination unit to determine a reference block to compart a frame of a reference layer of a location corresponding to an improvement block to compart a frame of an improvement layer; and a block structure reference unit to determine a limit depth of the improvement with reference to structure information in which the determined reference block is divided.
摘要翻译：公开了一种用于预测可伸缩视频的高速编码的层间结构的装置和方法。根据本发明的用于预测多层视频的层间结构的方法包括以下步骤：确定参考块以比较对应于改进块的位置的参考层的帧，以便将改进的帧层; 以及参考所确定的参考块被划分的结构信息来确定改进块的极限深度。根据本发明的用于多层视频的层间编码/解码的装置包括：相应的块确定单元，用于确定参考块以便比较对应于改进块的位置的参考层的帧以比较帧的改进层; 以及块结构参考单元，用于参考所确定的参考块被划分的结构信息来确定改进的极限深度。

64. 发明公开

KR1020140038322A 다중 계층 비디오를 위한 계층 간 필터링 장치 및 방법 审中-实审
标题翻译：用于多层视频的层间滤波的装置和方法
公开(公告)号：KR1020140038322A
公开(公告)日：2014-03-28
申请号：KR1020130111028
申请日：2013-09-16
申请人： 광운대학교 산학협력단
发明人： 남정학 , 유은경 , 심동규
IPC分类号： H04N19/50 , H04N19/00
CPC分类号： H04N19/53 , H04N19/117 , H04N19/31 , H04N19/59 , H04N19/80
摘要： Disclosed are a decoding apparatus for a multi-layer video, and an apparatus and a method for integer pixel filtering for inter-layers. The decoding apparatus can comprise: an inter-layer filter unit for generating an reference image by converting the size of a restoration image as much as a difference in spatial resolution between a restoration image of a reference layer and an input images of an enhancement layer, and by performing a filtering with respect to the restoration image; and a prediction unit for generating a prediction image of the enhancement layer at a location timely corresponding to the reference images by referring to the reference image. The method for integer pixel filtering for inter-layers can comprise the steps of: selecting at least one filter to perform a filtering with respect to a integer pixel of the restoration image; computing a filter coefficient for obtaining a filter coefficient with respect to a current integer pixel of the restoration image, which is an object to be filtered by at least one filter; and performing a filtering for generating a reference image by performing a filtering with respect to the current integer pixel of the restoration image based on the filter coefficient. [Reference numerals] (100) Integer pixel filtering apparatus; (110) Filter selecting unit; (120) Filter coefficient computing unit; (130) Filtering performing unit; (140) Filter coefficient transfer unit
摘要翻译：公开了一种用于多层视频的解码装置，以及用于层间整数像素滤波的装置和方法。解码装置可以包括：层间滤波器单元，用于通过将参考层的恢复图像与增强层的输入图像之间的空间分辨率的差异转换为恢复图像的大小来生成参考图像，并且通过对所述恢复图像执行滤波; 以及预测单元，用于通过参考参考图像在与参考图像相对应的位置处及时生成增强层的预测图像。用于层间整数像素滤波的方法可以包括以下步骤：选择至少一个滤波器以相对于恢复图像的整数像素执行滤波; 计算滤波器系数，以获得相对于由至少一个滤波器滤波的对象的恢复图像的当前整数像素的滤波器系数; 并且通过基于滤波器系数对恢复图像的当前整数像素执行滤波来执行用于生成参考图像的滤波。（附图标记）（100）整数像素滤波装置; （110）滤波器选择单元; （120）滤波器系数计算单元; （130）过滤执行单位; （140）滤波器系数转移单元

65. 发明公开

KR1020140038315A 다시점 영상을 부호화/복호화하는 장치 및 방법 审中-实审
标题翻译：用于编码/解码多视图的装置和方法
公开(公告)号：KR1020140038315A
公开(公告)日：2014-03-28
申请号：KR1020130110138
申请日：2013-09-13
申请人： 광운대학교 산학협력단
发明人： 유선미 , 남정학 , 심동규
IPC分类号： H04N19/50 , H04N13/00
CPC分类号： H04N19/50 , H04N13/271
摘要： Disclosed are an apparatus for encoding/decoding a video and a method for encoding/decoding a video which perform prediction between viewpoints in encoding/decoding for multi-view images. The apparatus for decoding a video using prediction between viewpoints comprises: a reference block sampling part which extracts the reference blocks of a reference point corresponding to the current blocks of an expansion point based on the disparity information of at least one neighbor block which is adjacent to the current blocks of the expansion point; and a prediction part which performs prediction on the current blocks of the expansion point by using prediction performance information on the reference blocks of the reference point. Accordingly, the present invention is able to improve the efficiency of encoding/decoding for multi-view images by effectively extracting the reference blocks of the reference point which is to be referenced by the current blocks of the expansion point. [Reference numerals] (210) Entropy decoding part; (210-1,210) Entropy decoding part; (220-1,220) Inverse quantization part; (230-1,230) Inversion part; (250-1,250) In-loop filter; (260-1,260) Frame memory; (270-1,270) Intra prediction part; (280-1,280) Motion compensation part; (29) Bit stream extraction part; (400) Reference block extraction part; (AA) Bit stream
摘要翻译：公开了一种用于对视频进行编码/解码的装置和用于对在多视图图像的编码/解码中的视点之间执行预测的视频进行编码/解码的方法。用于使用视点之间的预测对视频进行解码的装置包括：参考块采样部分，其基于与至少一个相邻块相邻的相邻块的视差信息提取与扩展点的当前块相对应的参考点的参考块当前块的扩展点; 以及预测部，其通过使用关于参考点的参考块的预测性能信息来对扩展点的当前块执行预测。因此，本发明能够通过有效地提取由扩展点的当前块参照的参考点的参考块，来提高多视点图像的编码/解码效率。（附图标记）（210）熵解码部分; （210-1,210）熵解码部分; （220-1,220）反量化部分; （230-1,230）反转部分; （250-1,250）环路滤波器; （260-1,260）帧存储器; （270-1,270）内部预测部分; （280-1,280）运动补偿部分; （29）比特流提取部分; （400）参考块提取部分; （AA）位流

66. 发明公开

KR1020120138426A 교차 삽입법을 이용한 향상된 움직임 백터 예측 방법 및 장치 无效
标题翻译：信号装置和交互式运动矢量预测的方法
公开(公告)号：KR1020120138426A
公开(公告)日：2012-12-26
申请号：KR1020110057897
申请日：2011-06-15
申请人： 광운대학교 산학협력단
发明人： 심동규 , 남정학 , 최효민
IPC分类号： H04N19/51 , H04N19/61
CPC分类号： H04N19/52 , H04N19/117 , H04N19/124 , H04N19/13 , H04N19/176 , H04N19/30 , H04N19/625 , H04N19/70
摘要： PURPOSE: A method for predicting an improved motion vector by using a cross insertion method and a device thereof are provided improve the efficiency of video encoding compression by performing more efficient motion vector prediction in the encoding of the motion vector with the cross insertion method. CONSTITUTION: An intra prediction unit(140) generates a prediction image block by using multidirectional intra prediction for a block of a screen image. An inter prediction unit(150) generates the prediction image block by using inter prediction through a motion information merge process and motion vector prediction improved by using a cross insertion method and a reference image of an image buffer for the block of the screen image. A reconstruction frame buffer(106) stores the whole reconstruction image block as a single screen. [Reference numerals] (106) Recontruction frame buffer; (140) Intra prediction unit; (150) Inter prediction unit; (AA) Input image; (BB) Encoding; (CC) In-loop filter
摘要翻译：目的：通过使用交叉插入方法及其装置来预测改进的运动矢量的方法通过使用交叉插入方法在运动矢量的编码中执行更有效的运动矢量预测来提高视频编码压缩的效率。构成：帧内预测单元（140）通过对屏幕图像的块进行多向帧内预测来生成预测图像块。帧间预测单元（150）通过使用运动信息合并处理的帧间预测和通过使用用于屏幕图像的块的图像缓冲器的交叉插入方法和参考图像而改进的运动矢量预测来生成预测图像块。重建帧缓冲器（106）将整个重建图像块存储为单个屏幕。（附图标记）（106）重新构造帧缓冲器; （140）内部预测单元; （150）帧间预测单元; （AA）输入图像; （BB）编码; （CC）环路滤波器

67. 发明公开

KR1020120070292A 에뮬레이션 방지 바이트 제거에 기초하여 순차적으로 비트 스트림을 파싱하는 장치 및 방법 审中-实审
标题翻译：基于去除模拟预防字节的进行性分段流的设备及其方法
公开(公告)号：KR1020120070292A
公开(公告)日：2012-06-29
申请号：KR1020100131791
申请日：2010-12-21
申请人： 삼성전자주식회사 , 광운대학교 산학협력단
发明人： 심동규 , 김두현 , 송준호 , 이시화 , 김도형 , 조현호 , 이진섭 , 서정한
IPC分类号： H04N19/423 , H04N19/42 , H04N19/44 , H04N19/70 , H04N21/434
CPC分类号： H04N19/42 , G06F9/30 , H04N19/44 , H04N19/70 , H04N19/423 , H04N19/439 , H04N21/434
摘要： PURPOSE: An apparatus and a method for parsing bit stream are provided to sequentially parse bit stream based on emulation prevention byte removal. CONSTITUTION: A first register buffer unit(110) stores bit stream. A second register buffer unit(140) stores a processing bit stream. An emulation prevention byte pattern is removed from the inputted bit stream. The process bit stream is generated. An output buffer selecting unit(150) selects the output of one or more register buffers based on the input of a buffer selection flag.
摘要翻译：目的：提供用于解析比特流的装置和方法，以便基于仿真预防字节移除来顺序解析比特流。构成：第一寄存器缓冲单元（110）存储比特流。第二寄存器缓冲器单元（140）存储处理位流。从输入的比特流中去除仿真预防字节模式。生成进程位流。输出缓冲器选择单元（150）基于缓冲器选择标志的输入来选择一个或多个寄存器缓冲器的输出。

68. 发明公开

KR1020120008321A 서브샘플링을 이용한 적응적 스캐닝 및 확장된 템플릿 매칭 방법 및 장치 无效
标题翻译：自适应扫描和扩展模板匹配设备和使用后备的方法
公开(公告)号：KR1020120008321A
公开(公告)日：2012-01-30
申请号：KR1020100069111
申请日：2010-07-16
申请人： 광운대학교 산학협력단
发明人： 심동규 , 남정학 , 조현호
IPC分类号： H04N19/59
CPC分类号： H04N19/59 , H04N19/124 , H04N19/129 , H04N19/625
摘要： PURPOSE: An adaptive scanning method using sub sampling, an extended template matching method, and a device thereof are provided to allocate an adaptive scanning order to coefficient values of a block unit which performs entropy encoding in a macro block. CONSTITUTION: A sub sampler(800) partitions an input image. A coefficient value pattern extracting unit generates a new scanning order by an encoded coefficient. A realignment unit(840) re-arranges the coefficient value of a current block using the new scanning order. An up-sampling unit(850) up-samples a discrete cosine transformed signal value and performs quantization. A template matching unit(860) predicts the current block by up-sampled block information.
摘要翻译：目的：提供一种使用子采样的自适应扫描方法，扩展模板匹配方法及其装置，以便在宏块中执行熵编码的块单元的系数值分配自适应扫描顺序。构成：子采样器（800）分割输入图像。系数值图案提取单元通过编码系数生成新的扫描顺序。重新排列单元（840）使用新的扫描顺序重新排列当前块的系数值。上采样单元（850）对离散余弦变换信号值进行上采样，并执行量化。模板匹配单元（860）通过上采样块信息预测当前块。

69. 发明公开

KR1020120008299A 화면 내 예측 블록의 영역별 특성을 고려한 적응적 필터 적용 장치 및 그 방법 无效
标题翻译：基于预测块区域特征的自适应滤波器和内部预测方法
公开(公告)号：KR1020120008299A
公开(公告)日：2012-01-30
申请号：KR1020100069081
申请日：2010-07-16
申请人： 광운대학교 산학협력단
发明人： 심동규 , 임웅
IPC分类号： H04N19/59
CPC分类号： H04N19/82 , H04N19/159
摘要： PURPOSE: An adaptive filtering apparatus considering a feature per each area in an intra prediction block and a method thereof are provided to use a selected filter based on a context in a generated prediction block, thereby minimizing the differential signal between an input image and a prediction block. CONSTITUTION: A context-based area separating unit(301) separates a block according to an analyzed context into at least one area with a similar feature. A filter determining unit(302) per an area determines the type of a filter to be used in each area. A filter applying unit(303) per an area uses the determined filter in a corresponding area. A filter application sample combining unit(304) generates a final prediction block by combining pixels of each area.
摘要翻译：目的：提供一种考虑帧内预测块中每个区域的特征的自适应滤波装置及其方法，以使用基于生成的预测块中的上下文的所选滤波器，从而最小化输入图像与预测之间的差分信号块。构成：基于上下文的区域分离单元（301）将根据分析的上下文的块与具有相似特征的至少一个区域分离。每个区域的滤波器确定单元（302）确定要在每个区域中使用的滤波器的类型。每个区域的过滤器施加单元（303）使用在相应区域中确定的过滤器。滤波器应用样本组合单元（304）通过组合每个区域的像素来生成最终预测块。

70. 发明公开

KR1020110014000A 영상 데이터의 디블록킹 필터링 장치 및 방법과 이를 이용한 영상 복호화 장치 및 방법 无效
标题翻译：消除图像数据和解码设备的装置和方法及其使用方法
公开(公告)号：KR1020110014000A
公开(公告)日：2011-02-10
申请号：KR1020090071730
申请日：2009-08-04
申请人： 삼성전자주식회사 , 광운대학교 산학협력단
发明人： 최웅일 , 조대성 , 남정학 , 심동규 , 조현호
IPC分类号： H04N5/21 , H04N7/24
CPC分类号： H04N19/117 , H04N19/157 , H04N19/176 , H04N19/61 , H04N19/86 , H04N5/21 , H04N7/24
摘要： PURPOSE: A deblocking filtering apparatus of image data, a method thereof, and an image decoding apparatus and method using the same are provided to selectively apply deblocking filtering according to a CBP parameter value and/or perform the application of a skip mode. CONSTITUTION: A deblocking filtering apparatus includes a parameter extraction unit(203), a determination unit(205) and a filter unit(207). The parameter extraction unit extracts the parameter information for plural unit blocks of a macro block from an inputted present frame. Based on the extracted parameter information, the determination unit determines deblocking filtering. According to the determination result, the filter unit performs the deblocking filtering for the macro block.
摘要翻译：目的：提供图像数据的去块滤波装置，方法及其使用方法，以及根据CBP参数值选择性地应用去块滤波和/或执行跳过模式的应用。构成：解块滤波装置包括参数提取单元（203），确定单元（205）和滤波器单元（207）。参数提取单元从输入的当前帧中提取宏块的多个单位块的参数信息。基于所提取的参数信息，确定单元确定去块滤波。根据确定结果，滤波器单元执行宏块的去块滤波。

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式