【作者】 田婷； 郭红星； 孙伟平； 周敬利；

【Author】 Tian Ting, Guo Hongxing, Sun Weiping, Zhou Jingli Department of Computer Science and Engineering, Huazhong University of Science & Technology, Wuhan, 430074, China

【机构】 华中科技大学计算机科学与工程系；

【摘要】 以视频编/解码理论为基础,提出一种在压缩域中,改变帧序列达到添加文字信息的方法——像素改变法。该方法在保持其它信息不变的情况下,修改宏块中需要改变的特定位置的像素的值。实验结果表明,对QCIF格式的标准测试序列,在压缩域进行的操作比在空间域进行的操作快4倍,对于I帧压缩视频,编辑后图像质量几乎没有变化;对于P帧压缩视频,随着帧数增加,嵌入文字质量有所下降,但对图像质量以及文字识别并无明显影响。此方法可用于在实时视频监控中进行文字信息嵌入工作。更多还原

【Abstract】 Growing with the development of the multimedia, the video sequences have become a new substrate of the information. The most specific character the video sequences have is that the original data is very large, and need to be compressed before transported, stored and processed. In practical situations, sometimes we need to do some edit on these compressed data. In traditional methods, we always decompressed the whole data and then edit them in the spatial domain, after that, we will compress the new data. Using these methods, the video quantity will be well, but they need a lot of calculations. So we need to exact some methods that can realize the edition in the compressed domain, such as embed the text information in compressed video streams. Basing on the video encoding/decoding theory and the study of the important methods around the world, the paper proposes a method to embed the text information in the video sequences, that is: pixel-transformation. In this method, we only change some pixels which are the components of the texts and so there need to do some derivation and get the correct values of the frame data. Though we have to decode a few numbers of macroblocks to do these derivations, the complexity is very low. There we use the bitstream that based on the H.263 compressed standard, and get the video sequences that changed in the compressed domain. Meanwhile, we get the video sequences that changed in the spatial domain. Experimental results show that the operations we do in the compressed domain is 4-times faster than in the spatial domain. And the image quality of the intra frames remains well. Though the quality of the embedded information will get worse with the growing of the numbers of the p-frames, it has no conspicuous impact on the quantity of the image and the text information still can be seen. The method can be used in the real time monitor systems.更多还原