At last, the adaptive vector quantization coding based on the DCT spectrum main orientation mentioned above is developed. 最后本文探讨了一种将基于结构特征的自适应分块与矢量量化相结合的编码方法。
Vector quantization ( VQ ) method becomes the main method used in real time visualization system because of its high compression ratio and fast decoding speed. 矢量量化因其压缩比大,解码简单效率高成为实时绘制系统中大规模纹理压缩主要使用的方法。
The computation of vector quantization which is mainly depending on the search of the nearest codeword is the main barrier of practicalization of vector quantization. 矢量量化的计算量主要在于搜索最近的码字,这也是矢量量化实用化的主要障碍。
Raises a method of window-based vector quantization for primitive division and a method based on RBF neural network for primitive classification directed to gain arc primitives from natural images and analysed the affect of 3 main parameters of it. 针对自然图象中圆弧基元的获取提出了窗口矢量化的分割方法和基于RBF神经网络的分类方法。详细分析了三个主要参数的作用。