期刊文献+

Freeman链码压缩算法的研究 被引量:18

Research on the Compression Algorithm for Freeman Chain Code
下载PDF
导出
摘要 文中提出一种新的链码编码方法 .在链码中的每个码是根据它与其前一个码之间的前进方向角度差来进行编码的 .统计结果表明 ,这样编码的每个码值的出现概率是很不同的 .角度差小的码值出现概率较大 ,而角度差大的码值出现概率较小 .这样结合霍夫曼 (Huffman)编码方法就可实现链码的压缩 .新链码与现有链码进行了比较 .比较结果显示新链码具有最小的平均码长 1.97位 /码 . This paper presents a new chain code. The new chain code is based on the Freeman chain code. It is educed from such an observation that when the chain of Freeman code moves along a digital curve or a border, each code element is often followed by the same or the neighbouring code of it. That is, when a chain moves in a direction defined by a code element, the probability that the direction indicated by the next code element is slightly changed is greater than that of sharply changed. We analysed 1050 curve or contour patterns in various applications. The statistic result for the probabilities of various angle differences between two contiguous elements in a chain is Angle differences: 0° ±45° ±90° ±135° 180° Probabilities:0.453 0.488 0.044 0.012 0.003 Let C 0, C 1, C 2, C 3, C 4, C 5, C 6 and C 7 represent respectively the angle change of 0°, 45°, -45°, 90°, -90°, 135°, -135° and 180°, accounts for 0.453, 0.244, 0.244, 0.022, 0.022, 0.006, 0.006 and 0.003.The symbols can be coded with Huffman method as C 0:0, C 1:10, C 2 :110, C 3:1110, C 4:11110, C 5:111110, C 6:1111110, C 7: 1111111. This is the new chain code. Each element except for the first in a chain is coded as a relative angle difference between it and the preceding element. The new chain code is compared with the existing ones and the result of the comparison shows that the new chain code has the least average bits per code of 1.97 bits/code.
作者 刘勇奎
出处 《计算机学报》 EI CSCD 北大核心 2001年第12期1294-1298,共5页 Chinese Journal of Computers
基金 国家自然科学基金 ( 6 940 30 0 4) 辽宁省科技基金 ( 0 14 5 0 8)资助
关键词 链码 霍夫曼编码 图像表示 链码压缩算法 FREEMAN链码 图像处理 chain code, Huffman code, image representation, chain code compression
  • 引文网络
  • 相关文献

参考文献4

  • 1Shih F Y,Pattern Recognition,1999年,32卷,7期,1203页
  • 2Yuan J,Pattern Recognition,1995年,28卷,5期,635页
  • 3Liu Y K,Comput Graph Forum,1993年,12卷,1期,21页
  • 4Liu Y K,Comput Graph Forum,1993年,12卷,1期,27页

同被引文献139

引证文献18

二级引证文献57

;
使用帮助 返回顶部