-
1.发明公开METHOD AND APPARATUS FOR IMAGE RETRIEVAL BASED ON A PROBABILISTIC NON-PARAMETRIC GEOMETRIC VERIFICATION 审中-公开方法和装置用于图像捕获不是基于概率几何参量检查
公开(公告)号：EP3073417A1
公开(公告)日：2016-09-28
申请号：EP15305457.2
申请日：2015-03-27
申请人： Thomson Licensing
CPC分类号： G06K9/4676
摘要： A technique for image retrieval is proposed based on a probabilistic non-parametric geometric verification which includes checking the intersection property within keypoint quadruplets of images. Additional checks may be performed for convexity and alignment of keypoints and inlier ratio. The technique yields better performance than standard techniques based on geometric models for cases of non-planar deformations as in 3D objects with larges changes in venture point, such as buildings, or lenses with large geometrical deformations, such as fisheyes. The technique may also be used to detect outliers which may be discarded, prior to the calculation of a geometric model, in order to improve the performance of such model.
摘要翻译： 一种用于图像检索技术基于包括检查图像的关键点四胞胎内的交叉点属性的概率性提出的非参数几何验证。 附加检查可以执行用于凸面和关键点和内点比率的对准。 该技术产生比基于几何模型的非平面变形的情况下，如3D与风险点拉尔变化，物体色彩标准技术更好的性能：如建筑物，或透镜与较大的几何变形，颜色：如鱼眼。 因此，可以使用该技术来检测可能被丢弃，一个几何模型的计算之前的异常值，以便改善搜索模型的性能。
-
2.发明授权Method for determining identifiers of segments that constitute an audiovisual document and receiver implementing the method 有权一种用于确定区段标识符方法以形成用于实现该方法的视听文件和接收器
-
3.发明公开Method for creating and accessing a menu for audio content without using a display 审中-公开一种用于生产和访问的菜单用于音频数据，而不使用屏方法
公开(公告)号：EP1437738A1
公开(公告)日：2004-07-14
申请号：EP03300170.2
申请日：2003-10-20
CPC分类号： G10H1/0008 , G10H2210/071 , G10H2210/081 , G10H2240/085 , G10H2240/135 , G11B27/105 , G11B27/28 , G11B27/34
摘要： A method for creating a menu (T) for audio content, e.g. music tracks, uses means (CL) for classifying the audio content into clusters (C1,...,C3) of similar tracks, the similarity referring to physical, perceptual and psychological features of the tracks. The method comprises a means (R) for automatic representative selection for clusters (C1,...,C3), and a means (X) for generating thumbnail representations of audio tracks. Said audio thumbnails are associated to the menu (T). Advantageously, no graphical or textual display is required for navigation, since the user may listen to an audio thumbnail and then enter a command, e.g. by pressing an appropriate button, for either listening to the related track or a similar track belonging to the same cluster, or listening to another type of music by selecting another thumbnail representing another cluster.
摘要翻译： 该方法涉及音频轨道的分类成群集，其中分类是基于轨道的特征参数进行。 轨道也被设置集群的代表被选择，其中选择是基于轨道的参数和集群的其它轨道的执行。 可再现的音频提取物是从代表轨道产生和提取物被关联到的菜单列表（T）。 因此独立claimsoft包括用于向用于创建或访问存储在存储单元上的音频内容的菜单装置。
-
公开(公告)号：EP1437738B1
公开(公告)日：2018-11-07
申请号：EP03300170.2
申请日：2003-10-20
申请人： Thomson Licensing
CPC分类号： G10H1/0008 , G10H2210/071 , G10H2210/081 , G10H2240/085 , G10H2240/135 , G11B27/105 , G11B27/28 , G11B27/34
摘要： The method involves classifying an audio track into clusters, where classification is performed based on a characteristic parameter of the track. The track that is being a representative for the cluster is selected, where selection is performed based on the parameter of track and of the other tracks of cluster. A reproducible audio extract is generated from the representative track and the extract is associated to a menu list (T). An independent claim is also included for an apparatus for creating or accessing a menu for audio content stored on a storage unit.
-
5.发明公开Method for determining identifiers of segments that constitute an audiovisual document and receiver implementing the method 有权一种用于确定区段标识符方法以形成用于实现该方法的视听文件和接收器
公开(公告)号：EP2581910A1
公开(公告)日：2013-04-17
申请号：EP12187428.3
申请日：2012-10-05
申请人： Thomson Licensing
摘要： The invention relates to a method for determining identifiers associated with segments of a document. Each segment consists of a series of individual elements such as images or sound sequences. Each segment of the document is subdivided into a determined number of portions comprising the same number of individual elements. An individual element is extracted from the most central portion of each segment and associated with the segment as identifier.
The invention also relates to the receiver capable of implementing the method.摘要翻译： 本发明涉及一种用于与文档的段相关联的确定性采矿标识符的方法。 如图像或声音序列：一系列单独的元件中的每个段besteht。 文档的每个段是细分成包含相同数量的单个元件的部分的确定性开采数。 单个元素从每个段的最中心部分提取和与所述段标识符相关联。 因此本发明涉及一种能够实现该方法的接收机。
-
6.发明公开Method to display colors using five primaries that allow colors to be metameric for most of the viewers 审中-公开一个使用5个基色，使其可以用于所有的颜色对于大多数观众的显示颜色的方法是metamerisch
公开(公告)号：EP2134089A1
公开(公告)日：2009-12-16
申请号：EP08305248.0
申请日：2008-06-11
申请人： Thomson Licensing
CPC分类号： H04N5/913 , H04N9/67 , H04N2005/91392
摘要： Two of the primaries G1', G2' are distributed in the green part of the visible spectrum, one G1' mainly distributed between 500 and 540 nm, the other one G2' mainly distributed between 540 nm and 585 nm. Thank to the specific spectral definition of the primaries B', G1', G2', R1' and R2', colors can be easily displayed that are metameric for most of the viewers, allowing notably anti-camcorder defeat in movie theatres without prejudice for the quality of display for most of the viewers.
摘要翻译： 初选G1的两个”，G2' 分布在可见光谱中，一个G1的绿色部分‘500和540nm之间主要分布，另一个G2’主要540纳米和585纳米之间分配。谢谢到特定光谱 初选B“ G1' ，G2”，R1和R2“颜色的定义可以很容易地显示根本是同色异谱对于大多数观众，允许在电影院特别是防摄像机的失利不影响显示的质量 大部分的观众。
-
7.发明授权Video coding with determination of coding order according to a minimum spanning tree 有权与确定的编码顺序的视频编码基于最小生成树
公开(公告)号：EP1613096B1
公开(公告)日：2008-07-16
申请号：EP05105563.0
申请日：2005-06-22
CPC分类号： H04N19/179 , H04N19/114 , H04N19/142 , H04N19/146 , H04N19/159 , H04N19/503 , H04N19/61 , H04N19/63
-
8.发明公开Method and apparatus for automatically calculating confidence values of data classification 审中-公开用于数据分类的安全值的自动计算的方法和装置
公开(公告)号：EP1413973A1
公开(公告)日：2004-04-28
申请号：EP02090360.5
申请日：2002-10-22
CPC分类号： G06K9/6217 , G06T7/41 , G06T2207/10016
摘要： E.g. in multimedia asset management systems, classification algorithms are used to classify multimedia data into different data classes, for instance such as indoor, landscape and city scenes, to enhance database (MDB) organisation. A problem is the confidence in classification performance. According to the invention, confidence measure values (c) about data classification results are calculated (MMDCL) in addition and can be included in user interaction. The confidence measure values are combined from a first part value ( c 1 ) that is related to the confidence in classification and from a second part value ( c 2 ) that is related to the confidence in said class models.
摘要翻译： 从多媒体类模型由...组成数据向量的计算置信度测量值被分配给每个输入的多媒体数据项。 所分配的值被存储在查找表中的所有其用于选择代表图像相对应的特定类别，并评估类含有由不同的类模型表示当前的多媒体数据和类之间的差异。 因此独立权利要求中包括了以下内容：（1）自动置信度量值计算装置; 和（2）的多媒体资产管理系统; 和（3）交互式图像浏览系统。
-
公开(公告)号：EP3096291B1
公开(公告)日：2017-12-20
申请号：EP15305735.1
申请日：2015-05-18
申请人： Thomson Licensing
IPC分类号： G06T7/20
CPC分类号： G11B27/34 , G06T7/215 , G06T7/246 , G06T2207/10016 , G06T2207/20104 , G06T2207/30241
摘要： The invention relates to a method for bounding an object in a video sequence F x,y,t . The method comprises obtaining a subset of pixels located in the object to annotate, in each frame of the video sequence. Spatio-temporal slicing is performed on the video sequence F x,y,t , centered on the obtained subsets of pixels, resulting in a first image F y,t obtained by an horizontal concatenation of first slices, comprising the obtained subsets of pixels, and resulting in a second image F x,t obtained by a vertical concatenation of second slices. A trajectory of the obtained subsets of pixels is displayed on both the first F y,t and second F x,t image. By using contour detection methods, a first and a second boundary are obtained on both the first F y,t and second F x,t image , around the trajectory of the obtained subsets of pixels. A bounding form around the object to annotate is obtained out of four points in each frame of the video sequence, wherein the coordinates of the four points of a frame t are obtained from the coordinates of the points located in the first and second boundary of the first and second image for that frame t. Advantageously, the bounding form is a rectangle drawn out of the four points, or an ellipse inscribed in that rectangle, or an ellipse comprising the four points.
-
10.发明公开METHOD AND DEVICE FOR PAIRWISE MATCHING OF FEATURES 审中-公开VERFAHREN UND VORRICHTUNG ZUR PAARWEISEN ANPASSUNG VON MERKMALEN
公开(公告)号：EP3096267A1
公开(公告)日：2016-11-23
申请号：EP15305747.6
申请日：2015-05-19
申请人： Thomson Licensing
CPC分类号： G06K9/6211 , G06F17/30
摘要： A method for pairwise matching features between two pieces of content is disclosed. It builds upon the idea that the best backward matching pairs generally belong to the k best forward matching pairs. Consequently, the pairwise matching method between a first set of features in a first piece of content and a second set of features in a second piece of content comprises a first step of forward matching resulting in forward matched pairs, wherein for each feature of the first set of features, a plurality of candidate forward matched pairs are selected, and a second step of backward matching applied in said plurality of candidate forward matched pairs for rejecting the forward matched pairs that do not belong to backward matched pairs.
摘要翻译： 公开了一种用于在两条内容之间成对匹配特征的方法。 它建立在这样的想法上，即最佳的后向匹配对通常属于k个最佳正向匹配对。 因此，第一内容中的第一组特征与第二内容中的第二组特征之间的成对匹配方法包括导致前向匹配对的正向匹配的第一步骤，其中对于第一内容中的每个特征 选择多个候选前向匹配对，并且在所述多个候选前向匹配对中应用反向匹配的第二步骤，用于拒绝不属于后向匹配对的前向匹配对。
-
-
-
-
-
-
-
-
-
搜索结果
30 条记录 (耗时 0.013 秒)
