会员体验
专利管家(专利管理)
工作空间(专利管理)
风险监控(情报监控)
数据分析(专利分析)
侵权分析(诉讼无效)
联系我们
交流群
官方交流:
QQ群: 891211   
微信请扫码    >>>
现在联系顾问~
热词
    • 4. 发明公开
    • Method for watermark detection using reference blocks comparison
    • 使用参考块比较的水印检测方法
    • EP2544143A1
    • 2013-01-09
    • EP12174263.9
    • 2012-06-29
    • Thomson Licensing
    • Baudry, SeverineRobert, AntoineLefebvre, Frederic
    • G06T1/00H04N7/26
    • G06T1/005G06T2201/0051G06T2201/0053G06T2201/0061G06T2201/0065G06T2201/0083
    • The invention relates to a method for decoding a watermark embedded in a target video comprising blocks of samples representative of the target video wherein watermark is embedded in an original video by modifying motion vectors of a compressed representation of the original video. The method comprises the steps of determining in the video a location of a watermarked block; determining a location for each reference block among a plurality of reference blocks wherein a reference block is associated to a watermark value among a plurality of watermark values; comparing the watermarked block with each reference block; and identifying watermark value for the watermarked block from the results of the plurality of comparison. The method is remarkable in that each reference block is a potential prediction block also located in the target video, for instance the pirated video, and in that the reference block that corresponds to the best prediction for the watermark block is selected. Thus the plurality of comparisons between the watermarked block and each reference block is performed among the same target video.
    • 本发明涉及一种用于对嵌入在目标视频中的水印进行解码的方法,该方法包括代表目标视频的样本块,其中通过修改原始视频的压缩表示的运动矢量来将水印嵌入到原始视频中。 该方法包括以下步骤:在视频中确定加水印块的位置; 确定多个参考块中的每个参考块的位置,其中参考块与多个水印值中的水印值相关联; 将加水印的块与每个参考块进行比较; 以及从多个比较的结果中识别加水印块的水印值。 该方法的显着之处在于每个参考块是也位于目标视频中的潜在预测块,例如盗版视频,并且选择对应于水印块的最佳预测的参考块。 因此,在相同的目标视频中执行加水印的块和每个参考块之间的多个比较。
    • 6. 发明公开
    • Method for watermark decoding
    • Verfahren zur Wasserzeichendecodierung
    • EP2682916A1
    • 2014-01-08
    • EP12305811.7
    • 2012-07-06
    • Thomson Licensing
    • Baudry, SeverineChupeau, Bertrand
    • G06T1/00
    • G06T1/0085G06T2201/0065G06T2201/0081
    • The invention relates to a method for decoding a watermark in a target video wherein a plurality of watermark decoding keys are associated with frame indices of a master video, the method comprising the steps of determining a frame index pair i,j wherein frame index i refers to a position of a frame in the master video and frame index j refers to a position of a frame in the target video; accessing a watermark decoding key Ki from the master frame index i; decoding watermark on frame j using the watermark decoding key Ki; computing a confidence score of a watermark decoding for the determined frame pair i,j; iterating previous step on i; matching a frame index in and a frame index j for which the confidence score of watermark decoding is maximum over the at least frame index i; and determining the watermark for the frame corresponding to frame index j in the target video by accessing key K i from frame index in from the matched frame index pair i n , j .
      In a preferred embodiment, the method is recursively applied to successive potential frame pairs and confidence score is input as a posteriori information in a registering algorithm. According to another aspect, the invention also relates to a device for implementing the method.
    • 本发明涉及一种用于对目标视频中的水印进行解码的方法,其中多个水印解码密钥与主视频的帧索引相关联,该方法包括以下步骤:确定帧索引i所指的帧索引对i,j 到主视频中的帧的位置,帧索引j是指目标视频中的帧的位置; 从主帧索引i访问水印解码密钥Ki; 使用水印解码密钥Ki对帧j上的水印解码; 计算确定的帧对i,j的水印解码的置信度分数; 迭代上一步我; 匹配帧索引in和帧索引j,其中水印解码的置信分数在至少帧索引i上是最大的; 以及通过从匹配的帧索引对i n,j的帧索引进入密钥K i来确定对应于目标视频中的帧索引j的帧的水印。 在优选实施例中,该方法被递归地应用于连续的潜在帧对,并且在注册算法中输入置信度作为后验信息。 根据另一方面,本发明还涉及一种用于实现该方法的装置。
    • 8. 发明公开
    • Method for watermarking free view video with blind watermark detection
    • Verfahren zur Wasserzeichenmarkierung frei ansehbarer视频mit Blindwasserzeichendetektion
    • EP2426636A1
    • 2012-03-07
    • EP11179468.1
    • 2011-08-31
    • Thomson Licensing
    • Doerr, GwenaelBaudry, SeverineFranco-Contreras, Javier
    • G06T1/00
    • G06T1/0085G06T1/005G06T2201/0065G06T2201/0083
    • The invention relates to a method for watermark embedding in multi views video comprising a plurality of views captured with a plurality of cameras comprising for each view a step of projecting the view (I) onto a domain invariant to the plurality of views and to intermediate synthesized view, which depends on the locations, orientations, and intrinsic parameters of the cameras, resulting into an invariant vector (i) ; a step of applying a watermarking algorithm to the invariant vector (i) resulting into a watermarked invariant vector (i w ) ; and a step of modifying the view (I) to obtain a watermarked view (I w ), wherein said watermarked view (I w ) when projected onto the invariant domain results into the watermarked invariant vector (i w ).
      The invention also relates to a method for watermark detection in a rendered view comprising a step of projecting the rendered view onto an invariant domain to the plurality of views resulting into an invariant rendered vector; a step of applying a watermark detection algorithm to the invariant rendered vector in order to assess the presence of a watermark.
    • 本发明涉及一种用于在多视图视频中嵌入水印的方法,包括用多个摄像机捕获的多个视图,所述多个摄像机包括用于每个视图的步骤,将视图(I)投影到对多个视图不变的域上,并将其中间合成 视图,其取决于相机的位置,方向和内在参数,导致不变向量(i); 将水印算法应用于不变量向量(i)的步骤,得到水印不变向量(i w); 以及修改视图(I)以获得水印视图(I w)的步骤,其中当投影到不变域上时,所述水印视图(I w)导致水印不变向量(i w)。 本发明还涉及一种用于在呈现视图中进行水印检测的方法,包括以下步骤:将呈现的视图投影到不变域上,导致产生不变的渲染矢量的多个视图; 为了评估水印的存在,将水印检测算法应用于不变渲染矢量的步骤。