huffman编码 原理与实现

分享一下我老师大神的人工智能教程!零基础,通俗易懂!http://blog.****.net/jiangjunshow

也欢迎大家转载本篇文章。分享知识,造福人民,实现我们中华民族伟大复兴!

               

哈夫曼算法原理


Wikipedia上面说的很清楚了,这里我就不再赘述,直接贴过来了。


1952年, David A. Huffman提出了一个不同的算法,这个算法可以为任何的可能性提供出一个理想的树。香农-范诺编码(Shanno-Fano)是从树的根节点到叶子节点所进行的的编码,哈夫曼编码算法却是从相反的方向,暨从叶子节点到根节点的方向编码的。

  1. 为每个符号建立一个叶子节点,并加上其相应的发生频率
  2. 当有一个以上的节点存在时,进行下列循环:
    1. 把这些节点作为带权值的二叉树的根节点,左右子树为空
    2. 选择两棵根结点权值最小的树作为左右子树构造一棵新的二叉树,且至新的二叉树的根结点的权值为其左右子树上根结点的权值之和。
    3. 把权值最小的两个根节点移除
    4. 将新的二叉树加入队列中.
  3. 最后剩下的节点暨为根节点,此时二叉树已经完成。

示例


huffman编码 原理与实现
Huffman Algorithm
huffman编码 原理与实现


符号 A B C D E
计数 15 7 6 6 5
概率 0.38461538 0.17948718 0.15384615 0.15384615 0.12820513

在这种情况下,D,E的最低频率和分配分别为0和1,分组结合概率的0.28205128。现在最低的一双是B和C,所以他们就分配0和1组合结合概率的0.33333333在一起。这使得BC和DE所以0和1的前面加上他们的代码和它们结合的概率最低。然后离开只是一个和BCDE,其中有前缀分别为0和1,然后结合。这使我们与一个单一的节点,我们的算法是完整的。


可得A代码的代码长度是1比特,其余字符是3比特。

字符 A B C D E
代码 0 100 101 110 111

    Entropy:huffman编码 原理与实现



Pseudo-code

 1:  begin 2:     count frequencies of single characters (source units) 3:     output(frequencies using Fibonacci Codes of degree 24:     sort them to non-decreasing sequence 5:     create a leaf node (character, frequency c, left son = NULL, right son = NULL)  6:      of the tree for each character and put nodes into queue F 7:     while (|F|>=2) do  8:      begin 9:        pop the first two nodes (u1, u2) with the lowest 10:         frequencies from sorted queue11:        create a node evaluated with sum of the chosen units, 12:         successors are chosen units (eps, c(u1)+c(u2), u1, u2)13:        insert new node into queue14:      end15:     node evaluate with way from root to leaf node (left son 1, right son 0)16:     create output from coded intput characters17end




哈夫曼算法实现



实现的时候我们用vector<bool>记录每个char的编码;用map<char,vector<bool>>表示整个字典;
就得到了下面的代码(下面有两个,一对一错):

先放出来这个错的,以示警戒

/************************************************************************//* File Name: Huffman.cpp*  @Function: Lossless Compression  @Author: Sophia Zhang  @Create Time: 2012-9-26 10:40  @Last Modify: 2012-9-26 11:10*//************************************************************************/#include"iostream"#include "queue"#include "map"#include "string"#include "iterator"#include "vector"#include "algorithm"using namespace std;#define NChar 8 //suppose use at most 8 bits to describe all symbols#define Nsymbols 1<<NChar //can describe 256 symbols totally (include a-z, A-Z)typedef vector<bool> Huff_code;//8 bit code of one charmap<char,Huff_code> Huff_Dic; //huffman coding dictionaryclass HTree{public : HTree* left; HTree* right; char ch; int weight; HTree(){left = right = NULL; weight=0;} HTree(HTree* l,HTree* r,int w,char c){left = l; right = r; weight=w; ch=c;} ~HTree(){delete left; delete right;} int Getweight(){return weight?weight:left->weight+right->weight;} bool Isleaf(){return !left && !right; } bool operator < (const HTree tr) const {  return tr.weight < weight; }};HTree* BuildTree(int *frequency){ priority_queue<HTree*> QTree; //1st level add characters for (int i=0;i<Nsymbols;i++) {  if(frequency[i])   QTree.push(new HTree(NULL,NULL,frequency[i],(char)i));    } //build while (QTree.size()>1) {  HTree* lc  = QTree.top();  QTree.pop();  HTree* rc = QTree.top();  QTree.pop();  HTree* parent = new HTree(lc,rc,parent->Getweight(),(char)256);  QTree.push(parent); } //return tree root return QTree.top();}void Huffman_Coding(HTree* root, Huff_code& curcode)if(root->Isleaf()) {  Huff_Dic[root->ch] = curcode;  return; } Huff_code& lcode = curcode; Huff_code& rcode = curcode; lcode.push_back(false); rcode.push_back(true); Huffman_Coding(root->left,lcode); Huffman_Coding(root->right,rcode);}int main()int freq[Nsymbols] = {0}; char *str = "this is the string need to be compressed"//statistic character frequency while (*str!='\0')  freq[*str++]++; //build tree HTree* r = BuildTree(freq); Huff_code nullcode; nullcode.clear(); Huffman_Coding(r,nullcode); for(map<char,Huff_code>::iterator it = Huff_Dic.begin(); it != Huff_Dic.end(); it++) {  cout<<(*it).first<<'\t';  Huff_code vec_code = (*it).second;  for (vector<bool>::iterator vit = vec_code.begin(); vit!=vec_code.end();vit++)  {   cout<<(*vit)<<endl;  } }}


上面这段代码,我运行出来不对。在调试的时候发现了一个问题,就是QTree优先队列中的排序出了问题,说来也是,上面的代码中,我重载小于号是对HTree object做的;而实际上我建树时用的是指针,那么优先级队列中元素为指针时该怎么办呢?

那我们将friend bool operator >(Node node1,Node node2)修改为friend bool operator >(Node* node1,Node* node2),也就是传递的是Node的指针行不行呢?


答案是不可以,因为根据c++primer中重载操作符中讲的“程序员只能为类类型或枚举类型的操作数定义重载操作符,在把操作符声明为类的成员时,至少有一个类或枚举类型的参数按照值或者引用的方式传递”,也就是说friend bool operator >(Node* node1,Node* node2)形参中都是指针类型的是不可以的。我们只能再建一个类,用其中的重载()操作符作为优先队列的比较函数。




就得到了下面正确的代码:


/************************************************************************//* File Name: Huffman.cpp*  @Function: Lossless Compression  @Author: Sophia Zhang  @Create Time: 2012-9-26 10:40  @Last Modify: 2012-9-26 12:10*//************************************************************************/#include"iostream"#include "queue"#include "map"#include "string"#include "iterator"#include "vector"#include "algorithm"using namespace std;#define NChar 8 //suppose use 8 bits to describe all symbols#define Nsymbols 1<<NChar //can describe 256 symbols totally (include a-z, A-Z)typedef vector<bool> Huff_code;//8 bit code of one charmap<char,Huff_code> Huff_Dic; //huffman coding dictionary/************************************************************************//* Tree Class elements:*2 child trees*character and frequency of current node*//************************************************************************/class HTree{public : HTree* left; HTree* right; char ch; int weight; HTree(){left = right = NULL; weight=0;ch ='\0';} HTree(HTree* l,HTree* r,int w,char c){left = l; right = r; weight=w; ch=c;} ~HTree(){delete left; delete right;} bool Isleaf(){return !left && !right; }};/************************************************************************//* prepare for pointer sorting*//*because we cannot use overloading in class HTree directly*//************************************************************************/class Compare_tree{publicbool operator () (HTree* t1, HTree* t2) {  return t1->weight> t2->weight; }};/************************************************************************//* use priority queue to build huffman tree*//************************************************************************/HTree* BuildTree(int *frequency){ priority_queue<HTree*,vector<HTree*>,Compare_tree> QTree; //1st level add characters for (int i=0;i<Nsymbols;i++) {  if(frequency[i])   QTree.push(new HTree(NULL,NULL,frequency[i],(char)i));    } //build while (QTree.size()>1) {  HTree* lc  = QTree.top();  QTree.pop();  HTree* rc = QTree.top();  QTree.pop();  HTree* parent = new HTree(lc,rc,lc->weight+rc->weight,(char)256);  QTree.push(parent); } //return tree root return QTree.top();}/************************************************************************//* Give Huffman Coding to the Huffman Tree*//************************************************************************/void Huffman_Coding(HTree* root, Huff_code& curcode)if(root->Isleaf()) {  Huff_Dic[root->ch] = curcode;  return; } Huff_code lcode = curcode; Huff_code rcode = curcode; lcode.push_back(false); rcode.push_back(true); Huffman_Coding(root->left,lcode); Huffman_Coding(root->right,rcode);}int main()int freq[Nsymbols] = {0}; char *str = "this is the string need to be compressed"//statistic character frequency while (*str!='\0')  freq[*str++]++; //build tree HTree* r = BuildTree(freq); Huff_code nullcode; nullcode.clear(); Huffman_Coding(r,nullcode); for(map<char,Huff_code>::iterator it = Huff_Dic.begin(); it != Huff_Dic.end(); it++) {  cout<<(*it).first<<'\t';  std::copy(it->second.begin(),it->second.end(),std::ostream_iterator<bool>(cout));  cout<<endl; }}




huffman编码 原理与实现



Reference:






关于Compression更多的学习资料将继续更新,敬请关注本博客和新浪微博Sophia_qing







           

给我老师的人工智能教程打call!http://blog.****.net/jiangjunshow

huffman编码 原理与实现