In this report, we propose an error correcting method for MPEG headers. This method uses a data hiding technique and a forward error correcting (FEC) code. At the proposed MPEG encoder, MPEG headers are encoded by FEC code. And then the resulting parity is extracted from data obtained after FEC coding. Next, the parity is hidden in the MPEG bitstream by means of data hiding technique. The reason for protecting MPEG headers is that the headers have the important information for video reproduction (picture size, quantization matrix, picture type, etc.). If errors occur in MPEG headers, the image quality is remarkably deteriorated in a decoded picture, and when the worst, decoding process can fail. The proposed decoder solves this problem by correcting the errors in MPEG headers, using the hidden parity. Thus, the bitstream output from the proposed encoder can, therefore, have error correction capability. In addition, preserving the standard MPEG structure, the bitstream can be decoded by the general MPEG decoders. Simulation results show the usefulness of the proposed method. Moreover, although MPEG-1 is the object of this research, the proposed method can be adapted also for MPEG-2 or MPEG-4.
展开▼
