A code conversion method from MPEG-4 SP to AVX video

Abstract

The invention relates to a MPEG-4 SP-to-AVS video code transformation method, which is characterized in that the method comprises the following steps: step one, a MPEG-4 SP code stream is decompressed fully to obtain the MPEG-4 SP data; step two, the MPEG-4 SP data is processed to obtain an image spatial domain sequence for specifying a space and the time resolution, an AVS image coding type, an AVS micro block coding mode and an AVS micro block motion vector; step three, an AVS code stream is obtained from the image spatial domain sequence for specifying the space and the time resolution, the AVS image coding type, the AVS micro block coding mode and the AVS micro block motion vector. A rapid MPEG-4 SP-to-AVS code transformation strategy based on a cascading structure in the invention can be suitable for the high fidelity real-time code transformation, the adaptability of the structure is strong, and the invention can meet a plurality of code transformation application requirements. The invention is in particularly suitable for the integration, and the concentrated browsing, transmission and memory of video resources under the heterogeneous condition of a cross-area cross-industry monitoring network at present.

Description

technical field [0001] The invention belongs to the technical field of multimedia communication and video coding, in particular to an MPEG-4 SP to AVS video transcoding method. Background technique [0002] Video transcoding technology is an effective means to solve the diversification of video resource access and transmission modes caused by different terminals and different network access methods under heterogeneous network conditions, and is the key to improving the efficiency of multimedia applications. The so-called video transcoding refers to the operation of converting a video stream in one encoding format into another encoding format. The "encoding format" here includes the standards or protocols (such as MPEG-2 / 4, MPEG-2 / 4, -4 Part10 AVC / H.264, H.26x, AVS, etc.), image spatial resolution (ie image size, such as CIF, QCIF, D1, etc.), image sequence temporal resolution (ie frame rate, such as 25fps, 30fps, etc. ), bit rate and coding tools (such as error-tolerant cod...

AI Technical Summary

Problems solved by technology

However, these existing transcoders have disadvantages: the open-loop transcoding structure completely omits the DCT / IDCT, so the operation is simple and the transcoding speed is fast, but the direct reweighting makes the motion vector mismatch, which eventually leads to a large prediction error, so the transcoded image quality is poor

The SDDT proposal and algorithm optimization are based on the assumption that DCT, IDCT, and MC are linear operations. However, in actual video coding, a large number of numerical interval clamping and rounding operations are used, which introduce non-negligible drift errors; in addition, SDDT has high control complexity and can only be used in rate-adaptive transcoding systems

[0005] Therefore, there is no complete implementation of MPEG-4 SP to AVS transcoding method, especially because AVS adopts rich macroblock coding modes, so it is necessary to design AVS fast mode selection algorithm based on transform domain

Images