公开(公告)号：KR1020050122496A

公开(公告)日：2005-12-29

申请号：KR1020040047616

申请日：2004-06-24

Applicant: 삼성전자주식회사

Inventor： 송병철

IPC: H04N19/577

CPC classification number: H04N19/573 ,  H04N19/176 ,  H04N19/513 ,  H04N19/577 ,  H04N19/61

Abstract: B 프레임의 부호화시 성능 저하 없이 연산량 및 메모리 밴드폭을 줄이기위한 B-픽쳐의 부호화 및 복호화 방법이 개시된다. 본 발명은 B-프레임의 현재 블록에 대해 제1 참조 프레임 내 소정의 탐색 영역 내에서 추정되는 제1 움직임 벡터에 대응하는 제1 참조 블록과, 제1 움직임 벡터로부터 유도된 제2 움직임 벡터에 대응하는 제2 참조 프레임 내 제 2 참조 블록과 제1 참조 블록의 평균 블록 중 하나의 블록을 현재 블록의 움직임 보상 블록으로 선정하는 과정, B 프레임의 움직임 보상 예측 모드 정보를 1비트로 생성하는 과정을 포함한다.

公开(公告)号：KR1020050118965A

公开(公告)日：2005-12-20

申请号：KR1020040044130

申请日：2004-06-15

Applicant: 삼성전자주식회사

Inventor： 송병철 ,  천강욱

IPC: H04N19/577

CPC classification number: H04N19/577 ,  H04N19/105 ,  H04N19/122 ,  H04N19/159 ,  H04N19/176

Abstract: 본 발명은 H.264 복호기에서의 B-프레임의 움직임 보상 방법 및 장치가 개시되어 있다. 본 발명은 임의의 B-매크로 블록내 한 서브 블록이 양 방향 예측 모드인 경우 현재 프레임과 참조 프레임간의 거리를 고려하여 한 방향의 참조 프레임에 존재하는 서브 블록을 선택하는 과정, 선택된 참조 서브 블록을 바탕으로 상기 B- 프레임의 움직임 보상을 수행하는 과정을 포함한다. 본 발명에 의하면 H.264 복호기에서 임의의 B-매크로 블록내 한 서브 블록이 양 방향 예측 모드인 경우 한 방향으로 단 한 번만의 MC만으로 B-프레임의 움직임 보상을 수행함으로써 메모리 대역폭을 줄일 수 있다.

公开(公告)号：KR1020050098214A

公开(公告)日：2005-10-11

申请号：KR1020050090686

申请日：2005-09-28

Applicant: 삼성전자주식회사

Inventor： 송병철 ,  천강욱

IPC: H04N19/136 ,  H04N19/124

CPC classification number: H04N19/126 ,  H04N19/13 ,  H04N19/14 ,  H04N19/176

Abstract: 본 발명은 개선된 동영상 부호화 방법에 관한 것으로서, 본 발명에 따른 복수개의 양자화 행렬을 사용하는 동영상 부호화 방법은 입력 영상의 특성에 기초하여 매크로블록 단위로 상기 복수개의 양자화 행렬 중 하나를 선택하는 단계와, 상기 입력 영상에 대해 영상 변환을 수행하는 단계와, 상기 영상 변환된 데이터에 대해, 상기 선택된 양자화 행렬을 사용하여 양자화를 수행하는 단계를 포함한다.

公开(公告)号：KR1020040107869A

公开(公告)日：2004-12-23

申请号：KR1020030038515

申请日：2003-06-14

Applicant: 삼성전자주식회사

Inventor： 송병철

IPC: H04N7/01

CPC classification number: H04N7/0112

Abstract: PURPOSE: A method and an apparatus for detecting a movie image according to grouping are provided to make a decision whether a received image signal is a movie signal, without setting a threshold value. CONSTITUTION: Some of similarity values are received in an image consisting of plural interlaced fields, wherein each similarity value is acquired by measuring a similarity of adjacent two fields of the same class in the image. The received similarity values are classified into the first group(410) and the second group(420). And then, the similarity values of the first and the second groups are converted to have different conversion values. It is decided whether the image is a movie image or not according to cycles of the converted values.

Abstract translation: 目的：提供一种根据分组检测电影图像的方法和装置，用于确定接收到的图像信号是否为电影信号，而不设置阈值。 构成：在由多个隔行场组成的图像中接收一些相似度值，其中通过测量图像中相同类的相邻两个场的相似度来获取每个相似度值。 所接收的相似度值被分类为第一组（410）和第二组（420）。 然后，将第一组和第二组的相似度值转换为具有不同的转换值。 根据转换的值的周期来决定图像是否是电影图像。

公开(公告)号：KR100441559B1

公开(公告)日：2004-07-23

申请号：KR1020020005857

申请日：2002-02-01

Applicant: 삼성전자주식회사

Inventor： 송병철

IPC: H04N7/01

Abstract: PURPOSE: A device and method for converting a scanning format are provided to adaptively change a scanning format of a macro block to reduce the quantity of calculations. CONSTITUTION: A scanning format converting device includes an intra-field interpolator(320), a motion compensator(340), and a data providing unit(310). The intra-field interpolator generates the second field from spatial-axis pixels of the first field of an input macro block to change a scanning format. The motion compensator converts the scanning format using the motion vector of the input macro block. The data providing unit adaptively decides a scanning format mode of the macro block on the basis of mode information and picture type of the macro block and the motion vector decoded, and then selectively provides the macro block to the intra-field interpolator and motion compensator.

Abstract translation: 目的：提供一种用于转换扫描格式的设备和方法，以自适应地改变宏块的扫描格式以减少计算量。 构成：扫描格式转换装置包括场内插值器（320），运动补偿器（340）和数据提供单元（310）。 场内插值器从输入宏块的第一场的空间轴像素产生第二场以改变扫描格式。 运动补偿器使用输入宏块的运动矢量来转换扫描格式。 数据提供单元基于宏块的模式信息和图片类型以及解码的运动矢量自适应地确定宏块的扫描格式模式，然后选择性地将宏块提供给场内插值器和运动补偿器。

公开(公告)号：KR1020040008359A

公开(公告)日：2004-01-31

申请号：KR1020020041985

申请日：2002-07-18

Applicant: 삼성전자주식회사

Inventor： 송병철 ,  천강욱

IPC: H04N19/53

CPC classification number: H04N19/105 ,  G06T7/207 ,  G06T7/238 ,  H04N19/112 ,  H04N19/176 ,  H04N19/53 ,  H04N19/61

Abstract: PURPOSE: A motion estimating method and apparatus using hierarchical motion vector search and a moving picture encoding system employing the method and apparatus are provided to minimize the quantity of calculations for estimating a field motion in a general hierarchical search method applied to MPEG-2. CONSTITUTION: A pre-processor(310) makes a current frame and a reference frame into a hierarchical structure through a low pass filtering and sub-sampling. A first motion estimator(320) searches a low-resolution frame formed by the pre-processor for at least initial search point having the minimum sum of absolute difference. A second motion estimator(330) carries out search in a high-resolution frame formed by the pre-processor with the initial search point as a reference motion vector of a frame and a field, and then estimates frame and field motion vectors having the minimum sum of absolute difference.

Abstract translation: 目的：提供使用分层运动矢量搜索的运动估计方法和装置以及采用该方法和装置的运动图像编码系统，以最小化应用于MPEG-2的通用分层搜索方法中用于估计场运动的计算量。 构成：预处理器（310）通过低通滤波和次采样使当前帧和参考帧成为分层结构。 第一运动估计器（320）搜索由预处理器形成的低分辨率帧至少具有绝对差的最小和的初始搜索点。 第二运动估计器（330）在由预处理器形成的高分辨率帧中以初始搜索点作为帧和场的参考运动矢量进行搜索，然后估计具有最小值的帧和场运动矢量 绝对差值之和。

公开(公告)号：KR100408294B1

公开(公告)日：2003-12-01

申请号：KR1020010054458

申请日：2001-09-05

Applicant: 삼성전자주식회사

Inventor： 송병철

IPC: H04N19/53 ,  H04N19/134

CPC classification number: H04N19/53 ,  H04N5/145

Abstract: A method and system for estimating a motion vector (MV) at high speed for low bit-rate coding is provided. The method and system for estimating a motion vector includes (a) forming sum pyramids for a current block and search blocks of a search region (b) setting an initial Cmin to C ({right arrow over (d ;lambda) (here, C ({right arrow over (dk;lambda) is an error measure at layer l which is determined by the following equation, and l and {right arrow over (d are 0 and (0,0), respectively:(c) if l is L, performing operation (d), if C ({right arrow over (d;lambda)>Cmin, performing operation (c) for a next candidate Motion Vector (MV) in the search region in a raster scanning method, after setting l to 0 (excluding (0,0) and {right arrow over (d ), and otherwise performing operation (c) after setting l to l+1; (d) if C ({right arrow over (d;lambda) ({right arrow over (d;lambda), if {right arrow over (d is the last candidate MV, performing operation (e), and otherwise performing operation (c) with the next candidate, after setting l to 0; (e) estimating the MV by determining the MV having Cmin as an optimal matcher. The motion vector estimation method is applied to a low bit-rate encoder, in which the ratio of motion vector is relatively high, so that motion vectors are estimated in real time.

Abstract translation: 提供了用于低比特率编码的高速估计运动矢量（MV）的方法和系统。 用于估计运动矢量的方法和系统包括：（a）形成当前块的和金字塔并搜索搜索区域的块;（b）将初始C min设置为C L; （在此，C 1（λ）是在层1处的误差度量，其由下面的等式确定，并且（和）（d是0和（0,0） ），分别为：（c）如果l是L，则执行操作（d），如果C（1） ）在将l设置为0之后（不包括（0,0）和（d）），并且在将l设置为1 + 1之后以其他方式执行操作（c）之后，在光栅扫描方法中， （d）如果C << L，则将Cmin更新为C L，将Cmin更新为C L， 候选人MV，执行操作（e），以及否则执行与下一个c的操作（c） 在将l设置为0之后， （e）通过确定具有Cmin作为最佳匹配器的MV来估计MV。 运动矢量估计方法应用于低比特率编码器，其中运动矢量的比率相对较高，从而实时估计运动矢量。

公开(公告)号：KR1020030072083A

公开(公告)日：2003-09-13

申请号：KR1020020011644

申请日：2002-03-05

Applicant: 삼성전자주식회사

Inventor： 송병철 ,  천강욱

IPC: H04N19/114 ,  H04N19/177

CPC classification number: G11B27/28 ,  G11B2220/20 ,  H04N9/8042 ,  H04N19/107 ,  H04N19/114 ,  H04N19/137 ,  H04N19/142 ,  H04N19/172 ,  H04N19/177 ,  H04N19/179 ,  H04N19/61 ,  H04N19/85 ,  H04N19/87

Abstract: PURPOSE: A method for encoding a moving picture and an apparatus appropriate for the same are provided to reduce a time required for playback by dividing GOP at a boundary frame and a representative frame of shots. CONSTITUTION: Input moving picture data is divided into GOP(Group Of Picture) units made up of I(Intra), B(Bidirectional), and P(Predicted) frames(S802). The input moving picture data is analyzed to detect a boundary between shots(S804). It is judged whether a current frame to be encoded is a boundary frame, by referring to the result of detecting the shot boundary(S806). If so, a GOP at a previous frame is terminated and a new GOP is started from the next frame(S808). If not, each frame is encoded according to the designated picture type(S810).

Abstract translation: 目的：提供一种用于对运动图像进行编码的方法和适用于该运动图像的装置，以通过在边界帧和代表性帧拍摄中划分GOP来减少播放所需的时间。 构成：输入运动图像数据被分成由I（Intra），B（Bidirectional）和P（Predicted）帧构成的GOP（Group Of Picture）单元（S802）。 分析输入运动图像数据以检测拍摄之间的边界（S804）。 通过参照检测拍摄边界的结果判断当前要编码的帧是否为边界帧（S806）。 如果是，则前一帧的GOP终止，并且从下一帧开始新的GOP（S808）。 如果不是，则根据指定的图像类型对每帧进行编码（S810）。

公开(公告)号：KR1020030069282A

公开(公告)日：2003-08-27

申请号：KR1020020008759

申请日：2002-02-19

Applicant: 삼성전자주식회사

Inventor： 송병철

IPC: H04N19/156

CPC classification number: H04N19/192 ,  H04N19/132 ,  H04N19/14 ,  H04N19/149 ,  H04N19/172 ,  H04N19/176 ,  H04N19/18 ,  H04N19/60 ,  H04N19/61

Abstract: PURPOSE: A method for encoding a moving picture having a fixed amount of operation, and an apparatus thereof are provided to regularly maintain the whole encoding computational complexity even if a random fast motion estimation method is used. CONSTITUTION: A DCT(Discrete Cosine Transform) operation amount calculator calculates the total computational complexity of a motion estimator in the previous frame for estimating a change to a target motion estimation computational complexity(510-520). The target motion estimation computational complexity is updated based on the estimated motion estimation amount change(530). A DCT skipping unit sets up a threshold for executing DCT skipping of the current frame by referring to the updated target motion estimation computational complexity(540-550). According to the set threshold, the DCT skip to the current frame is controlled(560-570).

Abstract translation: 目的：提供一种用于编码具有固定操作量的运动图像的方法及其装置，以便即使使用随机快速运动估计方法来定期地保持整个编码计算复杂度。 构成：DCT（离散余弦变换）运算量计算器计算用于估计对目标运动估计计算复杂度的改变的前一帧中的运动估计器的总计算复杂度（510-520）。 基于估计运动估计量变化来更新目标运动估计计算复杂度（530）。 DCT跳跃单元通过参考更新的目标运动估计计算复杂度建立用于执行当前帧的跳帧的阈值（540-550）。 根据设定的阈值，控制DCT跳到当前帧（560-570）。

公开(公告)号：KR1020150035315A

公开(公告)日：2015-04-06

申请号：KR1020130115699

申请日：2013-09-27

Applicant: 삼성전자주식회사 ,  인하대학교 산학협력단

Inventor： 김낙훈 ,  송병철 ,  배점한

IPC: H04N5/262

CPC classification number: H04N5/2355 ,  G06T5/003 ,  G06T5/50 ,  G06T2207/20208 ,  H04N5/23254 ,  H04N5/23267 ,  H04N5/35581

Abstract: 노출시간이동일한영상시퀀스를획득하고, 적어도하나의영상시퀀스와시간적으로각각대응되고, 노출시간이서로다른복수개의스틸(still) 영상을획득하고, 영상시퀀스를이용하여스틸영상으로부터하이다이나믹래인지(High dynamic range)를가지는정합영상을획득하는하이다이나믹레인지영상생성방법이개시된다.

Abstract translation: 公开了一种用于生成高动态范围图像的方法，其获取具有相同曝光时间的图像序列，时间上对应于至少一个图像序列，获取具有不同曝光时间的多个静止图像，并且获取具有 通过使用图像序列从静止图像的高动态范围。