-
公开(公告)号：KR1020160026281A
公开(公告)日：2016-03-09
申请号：KR1020140114476
申请日：2014-08-29
Applicant: 삼성전자주식회사 , 고려대학교 산학협력단
IPC: H04N5/217
Abstract: 다양한실시예에따른, 그리드패턴노이즈제거방법에있어서, 다수의디스플레이각각에따른그리드패턴노이즈에관한샘플영상들과전자장치가획득한입력영상을비교하여, 입력영상에포함된그리드패턴노이즈를추론하는동작과; 추론된그리드패턴노이즈를입력영상에서제거하는동작을포함할수 있다.
Abstract translation: 根据本发明的各种实施例，一种去除网格图案噪声的方法包括以下步骤：根据多个显示器中的每一个将与网格图案噪声相关联的样本图像与由电子设备获得的输入图像进行比较，以推断 包括在输入图像中的网格图案噪声; 并从输入图像中去除推断的网格图案噪声。
-
公开(公告)号：KR1020130061058A
公开(公告)日：2013-06-10
申请号：KR1020120125545
申请日：2012-11-07
Applicant: 고려대학교 산학협력단
IPC: G06T7/00 , H04N21/8549 , H04N5/91 , G11B27/02
CPC classification number: G06K9/00751 , G06K9/00288
Abstract: PURPOSE: An image summary method using visual features in images and a system thereof are provided to calculate significance for a key frame in the images by using the visual features in the images and to select the key frame based on the significance, thereby summarizing the images. CONSTITUTION: An image summary system detects one or more video shots from images and detects a key frame which shows the detected video shot(S210). The image summary system detects a face from the detected key frame and detects a character after grouping key frames which the face is detected(S220). The image summary system detects a key frame group and calculates significance for each key frame in the detected key frame group(S230). The image summary system selects and extracts the key frame corresponding to the number of requested key frame(S240). [Reference numerals] (AA) Start; (BB) End; (S210) Detect one or more video shots and detect a key frame which represents the detected video shot from images by using a low level feature vector; (S220) Detect a face for a character in the detected key frame, group the detected key frame with the face, and detect the character; (S230) Detect a key frame group including the key frames having a smaller distance between the key frames than a predetermined threshold value and calculate significance for each key frame in the detected key frame group; (S240) Select and extract the key frame corresponding to the number of requested key frame by a user in order of significance
Abstract translation: 目的：提供使用图像中的视觉特征及其系统的图像汇总方法，以通过使用图像中的视觉特征来计算图像中的关键帧的重要性，并且基于重要性来选择关键帧，从而总结图像 。 构成：图像汇总系统检测图像中的一个或多个视频图像，并检测显示检测到的视频拍摄的关键帧（S210）。 图像汇总系统从检测到的关键帧中检测出面部，并且在检测到面部的关键帧分组之后检测出角色（S220）。 图像汇总系统检测关键帧组并计算检测到的关键帧组中每个关键帧的重要性（S230）。 图像汇总系统选择并提取与所请求的关键帧数相对应的关键帧（S240）。 （附图标记）（AA）开始; （BB）结束; （S210）通过使用低级特征向量检测一个或多个视频镜头并检测表示从图像拍摄的视频镜头的关键帧; （S220）检测检测到的关键帧中的字符的面部，将检测到的关键帧与脸部对齐，并检测字符; （S230）检测包括关键帧之间的距离小于预定阈值的关键帧的关键帧组，并计算检测到的关键帧组中的每个关键帧的重要性; （S240）按用户的顺序选择并提取与请求的关键帧数相对应的关键帧
-
公开(公告)号：KR101341808B1
公开(公告)日：2013-12-17
申请号：KR1020120125545
申请日：2012-11-07
Applicant: 고려대학교 산학협력단
IPC: G06T7/00 , H04N21/8549 , H04N5/91 , G11B27/02
Abstract: 본발명은영상내 비주얼특징을이용한영상요약방법및 시스템에관한것으로, 보다구체적으로는영상으로부터저수준특징벡터를이용하여적어도하나의비디오샷을검출한후, 검출된비디오샷을대표하는키프레임을검출하는키프레임검출단계; 검출된상기키프레임내 얼굴을검출하고, 얼굴이검출된키프레임을군집화하여등장인물을검출하는등장인물검출단계; 등장인물이검출된상기키프레임내 등장인물의노출횟수, 상기얼굴이검출된키프레임의저수준특징벡터에기초하여키프레임간거리가기설정된임계값보다작은거리를갖는키프레임들을포함하는키프레임군집을검출하고, 검출된키프레임군집내 각키프레임에대하여중요도를연산하는중요도연산단계; 및사용자로부터요청받은키프레임수에따라, 중요도가높은순서별로해당하는키프레임을선별하여추출하는선별단계;를포함한다. 이러한구성에의해, 본발명의영상내 비주얼특징을이용한영상요약방법및 시스템은요약하고자하는영상내 등장인물, 키프레임의샷길이및 색상정보와같은비주얼특징을이용하여상기영상내 키프레임에대한중요도를연산한후, 연산한중요도에기초하여키프레임을선별함에따라, 영상을용이하게요약할수 있는효과가있다.
-
公开(公告)号：KR102203232B1
公开(公告)日：2021-01-14
申请号：KR1020140114476
申请日：2014-08-29
Applicant: 삼성전자주식회사 , 고려대학교 산학협력단
Abstract: 다양한실시예에따른, 그리드패턴노이즈제거방법에있어서, 다수의디스플레이각각에따른그리드패턴노이즈에관한샘플영상들과전자장치가획득한입력영상을비교하여, 입력영상에포함된그리드패턴노이즈를추론하는동작과; 추론된그리드패턴노이즈를입력영상에서제거하는동작을포함할수 있다.
-
-
-
Search Results
4
records
(took 0.02 seconds)
