对无线网络环境下视频图像的鲁棒性传输问题进行了研究，提出了一种增强视频图像传输差错恢复能力的新方法。在编码器端采用基于DCT(discrete cosine transform)矩阵系数的快速分类算法提取图像区域分类特征与再同步信息等构成指示信息，并用考虑量化误差后改进的DCT系数强制奇偶修改方法将其嵌入到图像码流中，用于解码器端视频图像序列的差错检测、再同步和差错恢复，以此增强基于H．263 的视频图像传输的鲁棒性。实验结果表明，此方法可以有效地改善标准编解码器的传输质量，不额外增加信源传输码字，计算成本较小。

This paper gives a thorough study on video robust transmission over the noisy channel in wireless network. A novel method to enhance the ability for error resilient video transmission is proposed. In order to obtain better reconstruction quality and enhanced error concealment performance at the decoder with a H.263 baseline function, the block classification and edge orientation information,bit-length value of macro block and other information about current block are organized together in binary form by the encoder as an indicator used for achieving detection,resynchronization and reconstruction of the errors and lost data in received video sequence. A fast classification algorithm based on calculated DCT domain matrix coefficients is employed in the proposed approach to extract numerical classification information and edge orientation value of image blocks. And all these information is imperceptibly embedded into image frame bit-stream by even-odd signaling forcibly of the DCT coefficients which has improved reduction of the visual loss and influence of image caused by modified DCT coefficients with a view of counting coefficient integral difference. Experiments and simulations on typical video sequences show that the proposed approach has a significant improvement in transmission performance against conventional techniques used in standard baseline H.263 encoder and decoder without changing the transmitted bit-stream syntax and increasing computation cost and complexity.