摘要： 使用了一种新的分片线性逼近算法。算法首先对极大极小这一分片线性函数的紧凑表示形式做了改进，然后以线性规划方法作为主要工具，以极小化拟合误差为准绳，同时确定全部参数；实现了分片线性函数对非线性样本数据的整体拟合；充分发挥了分片线性逼近的优势。在此基础上，提出了一种基于分片线性逼近的图象压缩编码方法。这种方法具有解压缩速度快的优点，与其它的图象压缩方法（例如ＤＣＴ）相结合，能够提高图象的压缩效率。

关键词: 图象压缩编码, 分片线性函数, 非线性逼近

Abstract: In this paper, we make use of a novel piecewise linear approximation algorithm in which the minmax representation of piecewise linear model is improved and then whole parameters is estimated simultaneously by minimizing fitting error. The algorithm takes full of the advantage of the piecewise linear approximation. Finally, we discuss a new image compression coding method that is based on this algorithm. Comparing with other image compression coding methods, this method is faster when decompressing. Specially, when combining with other image compression methods, such as DCT, it will be more efficient.

中图分类号: 

• TP391.41