-
公开(公告)号：KR1020140093189A
公开(公告)日：2014-07-25
申请号：KR1020140005198
申请日：2014-01-15
Applicant: 한국전자통신연구원 , 한양대학교 산학협력단
CPC classification number: G06K9/00275 , G06K9/6215 , G06K2009/485
Abstract: The present specification provides a method and an apparatus for face tracking robust to face rotation. The face tracking method of the present specification comprises the steps of: receiving an image including a user′s face; predicting an area having the high possibility for the face and the degree of rotation of the face within the image; detecting a facial area where the face exists from the image based on the predicted value; calculating the similarity of the facial area in a horizontal direction to obtain the degree of rotation of the face; and compensating the predicted value, based on the position of the face and the degree of rotation thereof, to extract the compensated facial area.
Abstract translation: 本说明书提供了一种用于面部跟踪的方法和装置，其适用于面向旋转。 本说明书的面部追踪方法包括以下步骤：接收包括用户脸部的图像; 预测图像中的面部的脸部和脸部旋转的可能性高的区域; 基于所述预测值从所述图像检测所述面部存在的面部区域; 计算面部区域在水平方向上的相似度以获得脸部的旋转度; 并且基于面部的位置和其旋转程度来补偿预测值，以提取补偿面部区域。
-
公开(公告)号：KR101409619B1
公开(公告)日：2014-07-02
申请号：KR1020140038594
申请日：2014-04-01
Applicant: 한국전자통신연구원 , 내셔널 유니버시티 오브 사이언시즈 앤드 테크놀로지
Abstract: 회귀 베지어 패치 분할에 기반하는 기하 보정 장치 및 방법이 개시된다. 본 발명의 일실시예에 따른 기하 보정 방법은 프로젝터가 투사 표면에 투사하는 검은 화면을 촬영한 제1 영상을 카메라로부터 수신하는 단계; 상기 프로젝터가 상기 투사 표면에 투사하는 미리 설정된 패턴을 촬영한 제2 영상을 상기 카메라로부터 수신하는 단계; 상기 제2 영상으로부터 상기 제1 영상을 빼서 제3 영상을 생성하는 단계; 및 상기 미리 설정된 패턴과 상기 제3 영상 사이의 왜곡을 보정하기 위하여, 상기 미리 설정된 패턴에 대하여 기하 보정을 수행하는 단계를 포함할 수 있다.
-
公开(公告)号：KR1020140077665A
公开(公告)日：2014-06-24
申请号：KR1020120146722
申请日：2012-12-14
Applicant: 한국전자통신연구원
IPC: H04N5/262
CPC classification number: H04N5/23238 , G02B5/08 , H04N5/23296
Abstract: According to an embodiment of the present invention, a multiple images photographing system includes: a mirror which is arranged on a two-dimensional plate and includes a curved outer surface; a plurality of image photographing devices which encloses the mirror on the two-dimensional plate, controls a view angle as the position is changed on the two-dimensional plate, and generates an image by photographing a predetermined object using the controlled view angle; and an image connection unit generating a panorama image by connecting the images generated by the image photographing devices.
Abstract translation: 根据本发明的实施例，多图像拍摄系统包括：反射镜，布置在二维平板上并包括弯曲的外表面; 在二维板上包围反射镜的多个图像拍摄装置控制二维平板上的位置改变的视角，并且通过使用受控视角拍摄预定物体来生成图像; 以及图像连接单元，通过连接由图像拍摄装置生成的图像来生成全景图像。
-
公开(公告)号：KR1020140067220A
公开(公告)日：2014-06-05
申请号：KR1020120134194
申请日：2012-11-26
Applicant: 한국전자통신연구원
IPC: H04N21/2343 , H04N21/242
CPC classification number: H04N21/242 , H04N21/2343
Abstract: The present invention relates to a method and an apparatus for transmitting and receiving a multi-stream in an interactive service, and provides a method for transmitting a second stream in synchronization with a first stream being transmitted by a server, and the method comprises steps of receiving, from a client, a second stream synchronization reproduction request message indicating a request for transmission of the second stream such that the second stream should be synchronized with the first stream and including an identifier of the first stream, checking the first stream corresponding to the second stream based on the identifier, and synchronizing the second stream with the first stream based on a reproduction point of the first stream; and transmitting the second stream to the client, starting from the reproduction point synchronized with the first stream. When multiple streams need to be reproduced in synchronization, the server matches a reproduction point of another stream with reference to any one stream and transmits the other stream to the client, thus, minimizing receiving of unnecessary stream data, a waste of a buffer, and the time for the client to wait for reproducing a stream.
Abstract translation: 本发明涉及一种用于在交互式服务中发送和接收多流的方法和装置，并且提供了一种与由服务器发送的第一流同步地发送第二流的方法，并且该方法包括以下步骤： 从客户端接收指示第二流的传输请求的第二流同步再现请求消息，使得第二流应与第一流同步并且包括第一流的标识符，检查对应于第一流的第一流 基于所述标识符的第二流，以及基于所述第一流的再现点使所述第二流与所述第一流同步; 以及从与第一流同步的再现点开始，将第二流发送到客户端。 当需要同时再现多个流时，服务器将参考任何一个流匹配另一个流的再现点，并将另一个流发送给客户端，从而最小化不必要的流数据的接收，缓冲区的浪费，以及 客户端等待再现流的时间。
-
公开(公告)号：KR1020140060090A
公开(公告)日：2014-05-19
申请号：KR1020120126760
申请日：2012-11-09
Applicant: 한국전자통신연구원
CPC classification number: G06F3/0233 , G06F3/018
Abstract: Disclosed in the present invention is a dynamic key button display method capable of shortening key input time by dynamically arranging keys according to how often each Korean alphabet is used and being differentiated for each user. According to the present invention, the dynamic key button display method includes: a step of extracting an initial consonant group from a first database and displaying the extracted initial consonant group; a step of selecting an initial consonant from the initial consonant group and displaying the selected initial consonant; a step of extracting a vowel candidate group applicable to the selected initial consonant from a second database and displaying vowels in the vowel candidate group in the descending order of use frequency how often each vowel is used together with the selected initial consonant in a way that the most frequently used vowel is placed the closest to the selected initial consonant; a step of selecting a vowel from the vowel group, displaying a letter which is the combination of the selected initial consonant and the selected vowel, and updating the use frequency of the selected vowel into the second database; a step of extracting an available final consonant group for the letter from a third database and displaying consonants in the final consonant group in the descending order of use frequency how often each consonant is used together with the letter in a way that the most frequently used consonant is placed the closest to the letter; and a step of selecting a final consonant from the final consonant group and updating the use frequency of the selected final consonant into the third database.
Abstract translation: 在本发明中公开了一种动态按键显示方法，其能够通过根据每个韩语字母表使用多少频率并且针对每个用户进行区分来动态地排列键来缩短键输入时间。 根据本发明，动态按键显示方法包括：从第一数据库提取初始辅音组并显示提取的初始辅音组的步骤; 从初始辅音组中选择初始辅音并显示所选择的初始辅音的步骤; 从第二数据库提取适用于所选择的初始辅音的元音候选组的步骤，并且以使用频率的降序的顺序在元音候选组中显示元音，每个元音的频率与所选择的初始辅音一起使用， 最常用的元音被放置在最接近所选择的初始辅音; 从元音组中选择元音的步骤，显示作为所选择的初始辅音和所选择的元音的组合的字母，以及将所选择的元音的使用频率更新为第二数据库; 从第三个数据库中提取可用的最终辅音组的步骤，按照使用频率的降序使辅音的使用次数与信件一起使用最常用的辅音的方式在辅音组中显示辅音 被放置在最靠近字母的位置; 以及从最终辅音组中选择最终辅音并将所选择的最终辅音的使用频率更新为第三数据库的步骤。
-
Patent Agency Ranking
公开(公告)号：KR1020140011964A
公开(公告)日：2014-01-29
申请号：KR1020130084582
申请日：2013-07-18
Applicant: 한국전자통신연구원
CPC classification number: F41H11/00 , G06K9/3233 , G06K2017/0045 , G06T5/50 , H04N5/23238
Abstract: A surveillance and reconnaissance method using a high-quality omnidirectional video is provided. The surveillance and reconnaissance method using a panoramic video comprises the following steps of: obtaining, by a panoramic video producing device, multiple images using multi-cameras; producing, by the panoramic video producing device, a panoramic video by stitching images which are adjacent to each other in a space-time domain among the multiple images; transmitting, by the panoramic video producing device, the panoramic video to a panoramic video playing device in real time; playing, by the panoramic video playing device, the panoramic video received from the panoramic video producing device; recognizing, by the panoramic video playing device, at least one target in the received panoramic video; and selecting, by the panoramic video playing device, a region of interest centered on the target, and tracking the target based on the selected region of interest. [Reference numerals] (110) Obtain multiple images using multi-cameras; (120) Produce a panoramic video by stitching images which are adjacent to each other in a space-time domain; (130) Transmit the produced panoramic video to a panoramic video playing device in real time; (140) Play the received panoramic video by the panoramic video playing device; (150) Recognize at least one target in the panoramic video; (160) Select a region of interest centered on the recognized target and track the target based on the region of interest; (AA) Start; (BB) End
Abstract translation: 提供使用高质量全向视频的监视和侦察方法。 使用全景视频的监视和侦察方法包括以下步骤：通过全景视频制作设备获得使用多摄像机的多个图像; 通过全景视频制作装置，通过拼接在多个图像之间的时空域彼此相邻的图像来产生全景视频; 将全景影像制作装置实时地将全景影像传送给全景影片播放装置; 通过全景视频播放设备播放从全景视频制作设备接收的全景视频; 通过全景视频播放设备识别所接收的全景视频中的至少一个目标; 以及通过全景视频播放设备选择以目标为中心的感兴趣区域，并且基于所选择的感兴趣区域来跟踪目标。 （附图标记）（110）使用多摄像机获取多个图像; （120）通过在时空域中拼接彼此相邻的图像来生成全景视频; （130）将生成的全景视频实时传输到全景视频播放设备; （140）通过全景视频播放装置播放所接收的全景视频; （150）识别全景视频中至少一个目标; （160）选择以识别的目标为中心的感兴趣区域，并根据感兴趣的区域跟踪目标; （AA）开始; （BB）结束
-
公开(公告)号：KR1020140009933A
公开(公告)日：2014-01-23
申请号：KR1020130082489
申请日：2013-07-12
Applicant: 한국전자통신연구원
IPC: G06F3/02
CPC classification number: G06F3/0233 , G06F3/04886
Abstract: The present invention relates to a key arrangement and input type of a virtual keyboard. The present invention provides a text input method using a virtual keyboard including a plurality of consonant and vowel groups. The text input method comprises the steps of: selecting one among consonant and vowel groups; and selecting one among consonants and vowels included in the selected consonant and vowel group. When the user uses the virtual keyboard, the text input method allows the user to select a consonant and vowel group and then select consonants and vowels included in the selected consonant and vowel group to reduce the number of consonants and vowels displayed on one screen, thereby relatively increasing the size of corresponding consonant and vowel keys and to reduce a typing error, thus improving text input speed.
Abstract translation: 本发明涉及虚拟键盘的键配置和输入类型。 本发明提供使用包括多个辅音和元音组的虚拟键盘的文本输入方法。 文本输入法包括以下步骤：选择辅音和元音组之一; 并选择包括在选定的辅音和元音组中的辅音和元音中的一个。 当用户使用虚拟键盘时，文本输入法允许用户选择辅音和元音组，然后选择所选辅音和元音组中包含的辅音和元音，以减少在一个屏幕上显示的辅音和元音的数量，从而 相应增加相应辅音和元音键的大小，减少打字错误，从而提高文字输入速度。
-
公开(公告)号：KR1020130068173A
公开(公告)日：2013-06-26
申请号：KR1020110134510
申请日：2011-12-14
Applicant: 한국전자통신연구원
CPC classification number: H04N5/23238
Abstract: PURPOSE: An apparatus for playing a spatial trick mode and a method thereof are provided to receive a video having a quality which is appropriate for a view navigation velocity by utilizing characteristics of visual recognition, thereby maintaining quality of experience while securing efficiency of bandwidth. CONSTITUTION: A panoramic video replay quality selector(103) determines a video quality hierarchy for replay based on a view navigation velocity that is detected by a view navigation velocity detector(102). A spatial-temporal trick mode-based video streaming receiver(104) requests for or receives video data of a corresponding view by considering a replay time according to the detected view navigation velocity. A visual recognition-based panoramic video replay unit(105) replays the received video data at the detected view navigation velocity. [Reference numerals] (100) Spatial trick mode playing apparatus; (101) Trick mode playing controller; (102) View navigation velocity detector; (103) Panoramic video replay quality selector; (104) Spatial-temporal trick mode-based video streaming receiver; (105) Visual recognition-based panoramic video replay unit
Abstract translation: 目的：提供一种用于播放空间特技模式的装置及其方法，用于通过利用视觉识别的特征来接收具有适合于视图导航速度的质量的视频，从而在保证带宽效率的同时保持体验质量。 构成：全景视频重播质量选择器（103）基于由视图导航速度检测器（102）检测到的视图导航速度来确定用于重放的视频质量层级。 基于空间 - 时间特技模式的视频流接收器（104）通过根据检测到的视图导航速度考虑重播时间来请求或接收对应视图的视频数据。 基于视觉识别的全景视频重播单元（105）以检测到的视图导航速度重放所接收的视频数据。 （附图标记）（100）空间特技模式播放装置; （101）模式演奏控制器; （102）查看导航速度检测器; （103）全景视频重播质量选择器; （104）基于空间特技模式的视频流接收机; （105）基于视觉识别的全景视频重放单元
-
公开(公告)号：KR1020130054868A
公开(公告)日：2013-05-27
申请号：KR1020110120519
申请日：2011-11-17
Applicant: 한국전자통신연구원 , 내셔널 유니버시티 오브 사이언시즈 앤드 테크놀로지
CPC classification number: H04N17/04 , H04N9/3185 , H04N9/3194
Abstract: PURPOSE: A geometrical correction device based on recursive Bezier patch sub-division and a method thereof are provided to display projected content on an uneven surface without distortion. CONSTITUTION: A geometrical correction device receives a first image that a projector photographs a black screen from a camera(501). The geometrical correction device receives a second image, which photographs a predetermined pattern, from the camera(502). The geometrical correction device generates a third image by deducting the first image from the second image(503). The geometrical correction device performs geometrical correction for the predetermined pattern in order to correct distortion between the predetermined pattern and the third image(504). [Reference numerals] (501) Receive a first image photographing a black screen from a camera; (502) Receive a second image photographing a predetermined pattern from the camera; (503) Generate a third image by deducting the first image from the second image; (504) Perform geometrical correction for the predetermined pattern; (AA) Start; (BB) End
Abstract translation: 目的：提供一种基于递归贝塞尔贴片子分割的几何校正装置及其方法，用于在不平坦表面上显示投影内容而不失真。 构成：几何校正装置接收投影仪从相机拍摄黑屏的第一图像（501）。 几何校正装置接收来自照相机（502）的预定图案的第二图像。 几何校正装置通过从第二图像（503）中扣除第一图像来生成第三图像。 几何校正装置对预定图案执行几何校正，以便校正预定图案和第三图像之间的失真（504）。 （附图标记）（501）从相机接收拍摄黑屏的第一图像; （502）从相机接收拍摄预定图案的第二图像; （503）通过从第二图像中扣除第一图像来生成第三图像; （504）对预定图案执行几何校正; （AA）开始; （BB）结束
-
公开(公告)号：KR1020120106550A
公开(公告)日：2012-09-26
申请号：KR1020120010044
申请日：2012-01-31
Applicant: 한국전자통신연구원
IPC: H04N19/147 , H04N19/164
CPC classification number: H04N21/6379 , H04N21/2662 , H04N21/43637 , H04N21/44209 , H04N21/6131 , H04N21/6377
Abstract: PURPOSE: A channel rate predicting method which uses signal strength information is provided to reduce the loss of video quality by using an LDPC(Low Density Parity Check) code. CONSTITUTION: A server predicts video channel rate by using estimated channel capacity. The server extracts a scalable video by using the predicted video channel rate. The sever performs channel coding of the extracted scalable video. The server transmits the channel-coded video to a terminal. The server minimizes video quality distortion by applying differential rate according to the property of a video frame through an LDPC code.
Abstract translation: 目的：提供使用信号强度信息的信道速率预测方法，以通过使用LDPC（低密度奇偶校验）码来减少视频质量的损失。 规定：服务器通过使用估计的信道容量来预测视频信道速率。 服务器通过使用预测的视频通道速率提取可缩放视频。 服务器执行提取的可伸缩视频的信道编码。 服务器将通道编码的视频发送到终端。 服务器通过利用LDPC码根据视频帧的属性应用差分速率来最小化视频质量失真。
-
-
-
-
-
-
-
-
-
Search Results
162
records
(took 0.029 seconds)
