视频编码量化失真估值算法研究

Research on Quantization Distortion Estimation Algorithm of Video Coding

针对量化失真估计算法存在的误差问题,对算法中的DCT系数分布模型进行了改进,提出了基于LAP分布模型的分段建模方法。将DCT系数的尾部数据进行分离,以主体数据的建模结果作为全体数据的分布模型。以均方误差作为量化失真的衡量准则,归纳并推到上述分布模型下的失真估计算法。采用限制量化步长的方法来评估失真估计算法的准确度。实验结果表明,采用LAP分布模型对DCT系数进行分段建模,不仅提高了模型的准确度,还简化了分布模型的复杂度;由此所产生的失真估计算法在复杂度和准确度方面具有很高的兼容性,与传统的LAP失真估计算法相比,准确度有所提高。

For the error problem of quantization distortion estimation algorithm,the discrete cosine transform(DTC)coefficient distribution model in the algorithm was improved,and a piecewise distribution model based on Laplacian(LAP)distribution model was proposed.The tail data of DCT coefficients were separated,taking the modeling results of the main data as the distribution model of the whole data.The mean square error was used to calculate quantization distortion,and the distortion estimation algorithm under this distribution model was generalized and deduced.The accuracy of the distortion estimation algorithm was evaluated by limiting the quantization step.Experimental results show that the proposed method not only improves the accuracy of the model,but also simplifies the complexity of the distributed model.The resulting distortion estimation algorithm has a high compatibility in complexity and accuracy,compared with the traditional LAP distortion estimation algorithm,the accuracy is improved.