公开(公告)号：US07782940B2

公开(公告)日：2010-08-24

申请号：US10633137

申请日：2003-08-01

申请人： Mohammad Athar Shah ,  Michael J. Horowitz

发明人： Mohammad Athar Shah ,  Michael J. Horowitz

IPC分类号： H04N7/18

CPC分类号： H04N7/15 ,  H04N19/105 ,  H04N19/132 ,  H04N19/134 ,  H04N19/136 ,  H04N19/137 ,  H04N19/159 ,  H04N19/176 ,  H04N19/196 ,  H04N19/197 ,  H04N19/198 ,  H04N19/507 ,  H04N19/587 ,  H04N19/61

摘要： Disclosed herein in one embodiment is an adaptive coding threshold algorithm for a videoconference system which alleviates problems due to the erroneous recoding of stationary areas of images caused by image noise. The algorithm, which is preferably implementable on a videoconference encoder, assesses the coding parameters for a current macroblock and the coding parameters for that same macroblock the last time it was coded. If for both macroblocks the coding was non-intra, with a [0,0] motion vector, and with a low quantization parameter, a determination is made that the current macroblock depicts a stationary image. Accordingly, the algorithm increases the coding threshold (T1) for the blocks within that macroblock, so that it becomes more difficult for those particular blocks to be recoded. This renders the block with stationary content less susceptible to erroneous recoding, while preserving system bandwidth and improving image quality. Additionally, related noise assessment techniques can be employed at the encoder or decoder without the adjustment of the coding threshold T1.

摘要翻译： 这里在一个实施例中公开了一种用于视频会议系统的自适应编码阈值算法，其减轻了由图像噪声引起的图像的静止区域的错误重新编码的问题。 优选地在视频会议编码器上可实现的算法评估当前宏块的编码参数以及上次编码时相同宏块的编码参数。 如果对于两个宏块，编码是非内部的，具有[0,0]运动矢量，并且在低量化参数下，确定当前宏块描绘静止图像。 因此，该算法增加了该宏块内的块的编码阈值（T1），使得这些特定块变得更加难以重新编码。 这使得具有固定内容的块不太容易被错误的重新编码，同时保持系统带宽并提高图像质量。 此外，相关的噪声评估技术可以在编码器或解码器处采用，而不需要调整编码阈值T1。