-
公开(公告)号：KR1020160054783A
公开(公告)日：2016-05-17
申请号：KR1020140154221
申请日：2014-11-07
Applicant: 삼성전자주식회사
IPC: G06T7/00
CPC classification number: G06K9/00281 , G06K9/00261
Abstract: 영상의랜드마크를결정하는방법및 장치가제공된다. 예를들어, 영상의랜드마크를결정하는장치는입력영상의관심영역으로부터검출된제1 후보랜드마크, 이전프레임으로부터추적된제2 후보랜드마크를기준으로설정된키랜드마크와비교하여, 키랜드마크와유사한랜드마크를최종랜드마크로결정할수 있다.
Abstract translation: 公开了一种用于从图像的兴趣区域及其装置确定地标的方法，其中所述装置包括：地标检测器，地标跟踪器和地标校正器。 例如，从图像的感兴趣区域确定地标的设备能够将被设置为标准的关键地标与从输入的兴趣区域检测到的第一候选地标进行比较 图像和从前一帧跟踪的第二候选地标; 从而确定与关键地标相似的地标作为最后的地标。 通过使用不同类型的属性信息，从本发明的图像的感兴趣区域确定地标的设备能够确定高度可靠的地标。
-
-
公开(公告)号：KR1020160029629A
公开(公告)日：2016-03-15
申请号：KR1020150001850
申请日：2015-01-07
Applicant: 삼성전자주식회사
IPC: G06K9/00
CPC classification number: G06K9/00241 , G06K9/00208 , G06K9/00268 , G06T7/00
Abstract: 얼굴인식방법및 장치가개시된다. 얼굴인식장치는사용자의얼굴영역이포함된 2D 입력영상을획득하고, 2D 입력영상으로부터얼굴의특징점을검출할수 있다. 얼굴인식장치는검출된특징점에기초하여미리저장된 3D 얼굴모델의포즈를조정하고, 조정된 3D 얼굴모델로부터 2D 투영영상을생성할수 있다. 얼굴인식장치는 2D 입력영상에나타난얼굴영역과 2D 투영영상에나타난얼굴영역을비교하여얼굴인식을수행하고, 얼굴인식결과를출력할수 있다.
Abstract translation: 公开了识别脸部的方法和装置。 面部识别装置能够：获得包括用户的面部区域的2D输入图像，以及从2D输入图像检测脸部的特征点; 基于检测到的特征点调整预先存储的3D脸部模型的姿态，以及从调整后的3D脸部模型生成2D投影图像; 通过将显示在2D输入图像中的面部区域与2D投影图像中显示的面部区域进行比较，然后输出面部识别结果来进行面部识别。
-
-
公开(公告)号：KR101527408B1
公开(公告)日：2015-06-17
申请号：KR1020080108747
申请日：2008-11-04
Applicant: 삼성전자주식회사
CPC classification number: G06K9/00221 , G06K9/00315 , G06T13/40
Abstract: 얼굴표정검출방법및 시스템을개시한다. 얼굴표정검출방법은입력된영상의얼굴기술자를추출하고상기얼굴기술자와포즈분류자를이용하여입력된영상의포즈를추정하고, 각사용자의특성에맞게개별화된데이터베이스를이용하여메쉬모델추출및 메쉬트래킹을수행함으로써, 얼굴표정을검출할수 있다.
-
Patent Agency Ranking
公开(公告)号：KR1020150046724A
公开(公告)日：2015-04-30
申请号：KR1020140118874
申请日：2014-09-05
Applicant: 삼성전자주식회사
CPC classification number: G06K9/00335 , G06K9/00234 , G06K9/6269 , G06K9/6277 , G06T7/97
Abstract: 얼굴트랙킹기술이개시된다. 일실시예에따르면, 입력영상에서검출된얼굴영역은서브영역들로분할되고, 서브영역들의가려짐확률들에기초하여입력영상에포함된얼굴이트랙킹된다.
Abstract translation: 披露了脸部跟踪技术。 根据一个实施例，在输入图像中检测到的脸部区域被分成子区域，并且基于子区域的闭合概率来跟踪包括在输入图像中的脸部。
-
公开(公告)号：KR1020150046718A
公开(公告)日：2015-04-30
申请号：KR1020140093315
申请日：2014-07-23
Applicant: 삼성전자주식회사
CPC classification number: G06K9/00335 , G06K9/00234 , G06K9/00268 , G06K9/6226 , G06T3/4007
Abstract: 얼굴트랙킹기술이개시된다. 일실시예에따르면, 입력영상에서검출된얼굴영역은서브영역들로분할되고, 서브영역들의가려짐확률들에기초하여입력영상에포함된얼굴이트랙킹된다.
Abstract translation: 公开了一种脸部跟踪技术。 根据实施例，将输入图像中检测到的面部区域划分为子区域，并且基于要覆盖的子区域的概率来跟踪包含在输入图像中的脸部。
-
公开(公告)号：KR1020150026354A
公开(公告)日：2015-03-11
申请号：KR1020130105083
申请日：2013-09-02
Applicant: 삼성전자주식회사
Abstract: The present invention relates to a method and an apparatus for matching medical images having different modalities. According to an aspect of the present invention, the method for matching medical images comprises the steps of: extracting a first anatomical subject including a point selected by a user in a first medical image obtained in real time; extracting a second anatomical subject in the second medical image, the second anatomical subject having an ultrasonic reflection characteristic different from that of the first anatomical subject; and matching the send medical image having a modality different from that of the first medical image with the first medical image by using the extracted first anatomical subject and second anatomical subject. The extracting of the first anatomical subject extracts the first anatomical subject by excluding an anatomical subject, which has the similar ultrasonic reflection characteristic as the first anatomical subject, appears a point other than the selected point.
Abstract translation: 本发明涉及一种用于匹配具有不同模态的医学图像的方法和装置。 根据本发明的一个方面，用于匹配医学图像的方法包括以下步骤：提取包括用户实时获取的第一医学图像中选择的点的第一解剖学对象; 在所述第二医学图像中提取第二解剖学对象，所述第二解剖学对象具有与所述第一解剖学对象不同的超声反射特征; 以及通过使用所提取的第一解剖学对象和第二解剖学对象，将具有与第一医学图像的形态不同的发送医学图像与第一医学图像进行匹配。 第一解剖学对象的提取通过排除具有与第一解剖学对象相似的超声反射特征的解剖学对象提取第一解剖学对象，出现除了所选择的点之外的点。
-
公开(公告)号：KR101434564B1
公开(公告)日：2014-08-29
申请号：KR1020080013061
申请日：2008-02-13
CPC classification number: G06K9/4619 , G06K9/6255 , H04N5/232 , H04N5/77 , H04N9/8227
Abstract: 촬영 정보 자동 추출 시스템 및 방법이 제공된다. 본 발명의 일 실시예에 따른 촬영 정보 자동 추출 시스템 프리뷰 영상 또는 촬영 영상을 입력 영상으로 획득하는 영상 입력부; 상기 입력 영상에 대한 촬영 정보를 추출하는 촬영 정보 추출부; 및 상기 추출된 촬영 정보를 이용하여 사용자의 촬영 패턴을 나타내는 촬영 코드를 생성하는 촬영 코드 생성부를 포함한다.
촬영 정보, 피사체 정보, 문맥 정보, 촬영 정보 추출, 촬영 패턴-
公开(公告)号：KR1020140096919A
公开(公告)日：2014-08-06
申请号：KR1020130010100
申请日：2013-01-29
Applicant: 삼성전자주식회사
CPC classification number: G06T3/0068 , A61B6/5288 , G06T7/337 , G06T2207/10072 , G06T2207/10132
Abstract: Disclosed are a method and a device for registering medical images having different modalities. The method for medical image registration according to an embodiment of the present invention comprises: a heterogeneous image registration step of registering a reference image among a plurality of first medical images photographed during a first respiratory cycle of a subject with a second medical image; a homogeneous image registration step of registering each of the remaining first medical images, which are acquired in different respiratory states from the reference image, with the reference image; a step of generating a set of registration images between the first medical images and the second medical images for the first respiratory cycle using the results of heterogeneous image registration and homogeneous image registration; and a step of acquiring a third medical image, which is photographed after the first respiratory cycle, in real time and detecting and outputting an image corresponding to the third medical image from the set of registration images.
Abstract translation: 公开了用于登记具有不同模态的医学图像的方法和装置。 根据本发明的实施例的用于医学图像配准的方法包括：异类图像登记步骤，在具有第二医学图像的被检体的第一呼吸循环期间拍摄的多个第一医疗图像之间登记参考图像; 将从参考图像以不同呼吸状态获取的剩余的第一医学图像的每一个与参考图像对齐的均匀图像配准步骤; 使用异构图像配准和均匀图像配准的结果，在第一医疗图像和第一呼吸周期的第二医疗图像之间生成一组注册图像的步骤; 以及获取在第一呼吸周期之后拍摄的第三医疗图像的步骤，并且从该组注册图像中检测并输出与第三医疗图像相对应的图像。
-
-
-
-
-
-
-
-
-
Search Results
195
records
(took 0.035 seconds)
