Demspter-Shafter theory application

Dempster-Shafer theory in decision making

Let us define

A= “not crowed at all”

B= “normal”

C=“ very crowed”

Then whole Environment E={A,B,C}. 


Vehicle       Message content      Message reputation score

V1                    C                                0.6       

V2                    C                                0.8      

V3                    A                                0.7      

(1) considering V1and V2(@ here means Unknown)

                                m1({C})=0.6           m1({@})=0.4

 m2({C})=0.8        {C}  0.48                 {C}   0.32

 m2({@})=0.2       {C}   0.12                {@}   0.08

Thus  m1+m2({C})=0.48+0.32+0.12=0.92         m1+m2({@})=0.08         

(2)  take V3 into consideration

                                       m1+m2 ({C})=0.92           m1+m2 ({@})=0.08

 m3({A})=0.7               NULL  0.644                          {A}   0.056

 m3({@})=0.3                {C}  0.276                             {@} 0.024

Which means m1+m2+m3({NULL})=0.644,  m1+m2+m3({A})=0.056

                        m1+m2+m3({C})=0.276,  m1+m2+m3({@})=0.024

(3) standardization because of Empty set

Because the m1+m2+m3({NULL})=0.644 should be 0, so after standard process:

m1+m2+m3({A})=0.056/(1-0.644)=0.157

m1+m2+m3({C})=0.276/(1-0.644)=0.775

m1+m2+m3({@})= 0.024/(1-0.644)=0.068

So, confidence interval :

A=[0.157, 0.157+0.068]=[0.157,0.225]

C=[0.775, 0.775+0.068]=[0.775,0.843]

Bad case for Dempster-Shafer theory
two doctor and one patient
1st doctor diagnoses the patient was attacted by tumor with 0.9 propability, and cancer with 0.1;
2nd doctor diagnoses the patient was attacted by concussion with 0.9 propability, and cancer with 0.1;
thus
                                                     m1({tumor})=0.9                  m1({cancer})=0.1

 m2({concussion})=0.9                {NULL}  0.81                   {NULL}   0.09

 m2({cancer})=0.1                       {NULL}   0.09                  {cancer}   0.01

=>    m1+m2({NULL})=0.81+0.09+0.09=0.99         m1+m2({tumor})=0.01  
       
       K= m1+m2({NULL})=0.99
       after standardization:
       m1+m2({tumor})=0.01/(1-K)=1
      Which would be greatly contrary with our intuition.
       So, when K is equal or very close to 1, we should just cancle the standardization

shared folder between virtualbox and host

1. Create a folder on host like this:

    F:\Shared

2.On VirtualBox Manager:

   Settings-Shared Folders-Add shared folder

   Folder Path: select the one you create above

   Folder Name: use the default one

   tick "Auto-mount"  and "Make permanent" and then save it.

3.Open terminal on Ubuntu on VirtualBox:

   sudo mount -t vboxsf Shared zhiguang/Win_Lin/

   # zhiguang/Win_Lin/ is the corresponding folder for Shared, you should create it before running this command

   then once you put files in  F:\Shared, you can share and edit them in zhiguang/Win_Lin/ immediately.
  

"The system is running in low-graphics mode" for Ubuntu on VM VirtualBox

When I installed Ubuntu Desktop 12.04 LTS into VirtualBox, it prompts "The system is running in low-graphics mode". I tried 6 soutions which solved the problem for other people but not me. Although I do not the exactly intrinsic reason, the following command finally served me well:

sudo apt-get update
sudo apt-get upgrade

sudo reboot

I hope it can also help you.

Full screen for ubuntu on VM VirtualBox

I have VM VirtualBox on Win 7, and Ubuntu on VM VirtualBox. By default, the ubuntu could not be displayed as fullscreen. But there is way to make it seem as the ubuntu os is directly installed on your PC hardisk.

1. Installing Guest Additions for full screen

   [Device]--->[Install Guest Addtions]

2. After some configure by the machine, you can open the Guest Additions folder which should be on your Desktop in Ubuntu with name like this: VBOXADDITIONS_#

3. Open terminal, and cd into this folder VBOXADDITIONS_# by:

cd  /media

cd  VBOXADDITIONS_# 

sudo sh VboxLinuxAdditions.run

4. Restart the ubuntu.
you can see the ubuntu is enlarged filling up the whole VirtualBox. However the VirtualBox window frame still exists, you can implement the real full screen by HOST+F,  which would make  the VirtualBox window frame  disappear. Another  HOST+F will bring the window frame back again. By default in Ubuntu, the HOST key is the Right Ctrl

ANSI C (2)

12. recursion

#include <stdio.h>

void binary_to_ascii( unsigned int value)

{

   unsigned int quotient;

   quotient=value/10;

   if(quotient!=0)

      binary_to_ascii(quotient);//recursion

   putchar(value%10+'0');



}



value=4267  quotient=?

value=4267  quotient=426

value=426  quotient=42

value=42  quotient=4

value=4  quotient=0

value=42  quotient=4

value=426 quotient=42

value=4267  quotient=426

13.variable parameter

#include <stdarg.h>

float average(int n_values,...)

{

  va_list var_arg;

  int count;

 float sum=0;



 var_start(var_arg,n_values);

for(count=0; count<n_values;count+=1)

{

   sum+=va_arg(var_arg, int);

}

va_end(var_arg);

return sum/n_values;

}

14. array

//(1)

int array[10];

2[array]  ==  *(2+(array)) ==array[2];



//(2)

int array[10], a;

for(a=0;a<10;a=a+1)

     array[a]=0;



int array[10], *ap;

for(ap=array; ap<array+10;ap++)  // more efficient here than array

   *ap=0;

// when   accessing the array elements by a fixed increment,  using pointer would more efficient than that of array index;



//(3)

#include <stdio.h>

#include <stdlib.h>

void my_parameter1(char string[]);

void my_parameter2(char *string);

int main()

{

    char arr[]="zhiguang";

    char *arr1="zhiguang";

    printf("%d\n",sizeof(arr));//9--array: 'z'  'h' ... 'g','\0'

    printf("%d\n",strlen(arr));//8--array:  'z'  'h' ... 'g'

    printf("%d\n",sizeof(arr1));//4--pointer size

    printf("%d\n",strlen(arr1));//8-- the content length the pointer points to

    my_parameter1(arr);//4--pointer size, it is not 9, because it only pass the array as pointer in fact

    my_parameter1(arr1);//4--pointer size

    my_parameter2(arr);//4--pointer size

    my_parameter2(arr1);//4--pointer size

    return 0;

}

void my_parameter1(char string[])

{

   printf("%d\n",sizeof(string));

}

void my_parameter2(char *string)

{

   printf("%d\n",sizeof(string));

}



//(4)

    char arr[]="zhiguang";

    char *arr1="zhiguang";//constant

    arr[2]='w';// ok

    arr1[2]='w';//error, because it is constant

    *(arr1+2)='w';//error, because it is constant



//(5)

int matrix[3][10];

matrix;//pointer, points to the first whole row, 10 elements

matrix+1;//pointer, points to the second whole row, 10 elements

*(matrix+1);// it is an array, also the points to the first element in second row;

*(matrix+1)+5;//pointer, points to matrix[1][5]

*(*(matrix+1)+5);//value, content of matrix[1][5]

*(matrix[1]+5);//value, content of matrix[1][5]

matrix[4,3]==matrix[3];//because of ','



//(6)

int vector[10], *vp=vector;//ok

int matrix[3][10], *mp=matrix;//error;

int (*p)[10]=matrix;//ok, p is a pointer first, points to an array with 10 elements;

int *pi=&matrix[0][0];//ok

int *pi=matrix[0];//ok



//(7)

// if the parameter mat is 2-D array, you can declare the function as:

void fun(int (*mat)[10]);//ok

void fun(mat[][10]);//ok

//however, the following is wrong

void func2(int **mat);//error



//(8)

char  *keyword[5];// an array, each element is a pointer;

char const *keyword[]={"do","for","while"};

// the best way to determine the length of this array is:

sizeof(keyword)/sizeof(keyword[0]);



//(9)
void function0(int);
void function1(int);
void function2(int);
int main(void)
{
   void (*f[3])(int) = {function0,function1,function2};

//f is first an array, each element is a pointer, pointing to a func with an int parameter

   return 0;
}

15. string

//strlen returns an unsigned in, so

if(strlen(x)>strlen(y)) //always right

if(strlen(x)-strlen(y)>=0) //sometimes, error

if(strlen(x)>=10)//always right

if(strlen(x)-10>=0) //sometimes, error

16. void *

//any type pointer could be converted to void*, and that is why

void  *memcpy(void *dst, void const *src, size_t length);

void  *memmove(void *dst, void const *src, size_t length);

void  *memcmp(void *a, void const *b, size_t length);

//...

17. structure

//(1)

struct{

   int a;

   int b;

}x;



struct SIMPLE{

   int a;

   int b;

};

struct SIMPLE x;



typedef struct {

   int a;

   int b;

} Simple;

Simple x;



//(2)

typedef struct{

    int a;

    struct SELF_REF3 *b;//error, there is no SELF_REF3 yet;

    int c;

} SELF_REF3; 



typedef struct SELF_REF3_TAG{

    int a;

    struct SELF_REF3_TAG *b;//OK, there is SELF_REF3_TAG already;

    int c;

} SELF_REF3_TAG; 



//(3)

// the following is error

struct A{

struct B *partner;

};

struct B{

struct A *partner;

}

//the following is correct

struct B;// declare first, even if it is not complete

struct A{

struct B *partner;

};

struct B{

struct A *partner;

18. memory

//(1)

void *malloc(size_t size);

void free(void *pointer);

// if malloc failed, it would return an NULL



//(2)

int *pi;

pi=malloc(25*sizeof(int));

if(pi==NULL)

{

   printf("out of memory!\n");

   exit(1);

}



//(3)

// the pointer passed to free must be generated by malloc, calloc, or realloc, not others;



//(4)

// simplest way of memory leak

for(int i=0; i<10;i++)

    int *p=malloc(25*sizeof(int));

ANSI C (1)

1. logically delete the command

# if 0

   statements

#endif

2. exit

exit(EXIT_SUCCESS);// stdlib.h

exit(EXIT_FAILURE);// stdlib.h

3.array variable as parameter
array variable, as function parameter, will pass the reference to the function, other kinds of variables will pass the values to the function
4.NUL  NULL

NUL=='\0' ;// the ASCII character 

NULL==0; // the pointer

5. read

#define MAX_INPUT 1000

char input[MAX_INPUT];

int ch;

while(gets(input)!=NULL)//read the line

{

}

while((ch=getchar())!=EOF&&ch!='\n')//read the character 

{

}

6. standard structure

#include <stdlib.h>

int main(void)

{

   return 0;

}

7.typedef  define

typedef char   *ptr_to_char;

define  p_ptr_to_char char*;

8 const pointer

int const *pci ; // you can modify pci, but not *pci

int *const pci;  // you can modify *pci, but not pci;

9 switch

switch(command)

{

  case 'A':

      add_entry();

      break;

  case 'B':

      delete_entry();

     // break;

  case 'C':

      edit_entry();

      break;

}

if command=='B', then delete_entry()  and  edit_entry() would be implemented, because there is no 'break' for case'B'. So pay attention to all the 'break'.
10 sizeof
it is an operator, not function

sizeof(int); // return the length by bytes

sizeof x;

11. pointer to string

//  string length

#include <stdlib.h>

size_t strlen(char *string)

{

   int length=0;

   while(*string++ !='\0')

          length+=1;

   return length;

}

// find the char in strings

#include <stdio.h>

#include<assert.h>

#define TRUE 1

#define FALSE 0

int find_char(char **strings, char value)

{

   assert(strings!=NULL)



   while(*strings !=NULL)

  {

       while(**strings!='\0') 

       {

            if(*(*strings)++==value)

                return TRUE;

       }

       strings++;

  }

   return FALSE;

}

Java basic

1. in one .java file, there could be as many class as you wish, but there should be at most 1 public class, and this class name should be the same with the file name.

// B.java

class A

{
}

public class B

{
}

class C

{
}

2. for boolean, there are only true and false, no -1, 0,1
in C++, you can write like this:

while(1)

{
}

but in Java, while(1) is wrong, and you can only use:

while (true)

{
}

3. define array

//-1

int num[];

num=new int[3]; //ok;

int num[]=new int[3];//ok

int num[]={1,2,3};//ok

int num[]=new int[]{1,2,3};//ok

int num[]=new int[3]{1,2,3};//error

int num[3]; //error



//-2

int [][] num;

num=new int[2][];

num[0]=new int[3]; //3 elements

num[1]=new int[4]; //4 elements

4. for loop

for(int i=0; i<5; i++)

{
   int a=i; //ok

}

int b=i;// error, i is not defined out of the for loop

5. print hex number

  int hx=0xffffffff;
  int hc=hx<<2;
  System.out.println(Integer.toHexString(hc));

6. default value

int   0;

char '\0'

boolean false

object  null

7.  same name for parameter and member

class Point

{
    int x,y;

    void func1(int x, int y)

   {

      x=x; //parameter to parameter, not what we want

      y=y; //parameter to parameter, not what we want

   }

  

    void func2(int x, int y)

   {

      this.x=x; //parameter x to member x

      this.y=y; //parameter y to member y

   }



}

8. call another construction method

 class Point

{
    int x,y;

    Point(int a, int b)

  {

     x=a;

     y=b;

  }



  Point()

 {

    this(3,4);// call another construct method, and should be the first statement

 }

}

9. static
static methods could call static methods and access the static variable, could not call the non-static methods or access variable;
while  non-static methods  could call static methods and access to static variable.
10.. final variable

In most cases, the final variable should be initialized when it is defined.

class Point

{

    final double PI=3.1415926;

}



one exception:

class Point

{

    final double PI;

    Point()

  {

     PI=3.1415926; 

  }

}

however, if the final variable is also static, you must initialize it when defining it:

static final double PI=3.1415926;

11. call parent method

public class MyAnimal
{
   MyAnimal()
  {
       System.out.println("I am Animal! without parameter");
  }

     MyAnimal(int param)
  {
       System.out.println("I am Animal! with parameter");
  }

}


 class MyFish extends MyAnimal
{

      public MyFish()
   {
      // in this place, super()is called even if you do not write it;

     // so if there is no MyAnimal() at all, error would happen.

     // in that case, you could use  super(0) if  MyAnimal(int param) is defined

     // or you could redifine the MyAnimal() 

      System.out.println("I am fish! without parameter");
   }

    public MyFish(int param)
   {
      //in this place, still super(),not super(0), is called even if you do not write it;

     // to call MyAnimal(int param), should write super(0) in this place

     // 0 represents the parameter, you can use anyone you like
      System.out.println("I am fish! with parameter!");
 
 } 

    public void breathe()
    {
          super. breathe(); // use this call  parent  breathe();

          System.out.println("Fish breathe!");

    }

}

12. polymorphic

class Parent

{

    void fun1()

  {

     System.out.println(" Parent 1");

  }

   void fun2()

  {

     System.out.println(" Parent 2");

  }

}



class Child extends Parent

{

  

   void fun1()

  {

     System.out.println(" Child 1");

  }  

  public static void main(String[] args)

 {

       Child cd=new Child();

       Parent pt=cd;//Ok

       pt.fun1();    // result: Child 1, because of new Child();

       pt.fun2();    // result: Parent 2, because Child dose not override fun2()



       Parent pt1=new Parent();//Ok

       Child cd1=pt1;//error, can not convert from Parent to Child

 }

}

read and write file in Java

import java.io.BufferedReader;
import java.io.BufferedWriter;
import java.io.FileReader;
import java.io.FileWriter;
import java.io.IOException;



public class Read_Write {

 public static void main(String[] arg) throws IOException{
 String fileName = "data/input/tweet_2012_06";
 BufferedReader in = new BufferedReader(new FileReader(fileName));
 String str;

 FileWriter fstream = new FileWriter("data/output_tweet_2012_06.txt");
 BufferedWriter out = new BufferedWriter(fstream);

 int counter = 0;
 while ((str = in.readLine())!= null){
  System.out.println(counter+"  " + str);
  out.write(str);
  counter++;

 }
    System.out.println("counter is "+counter);
  out.close();
 in.close();
 }

}

SendMessage from modaless dialog to the view in SDI without document/view structure support--another method

I create one SDI without document/view structure support, i name the project as ImagePro, then I can get the following files:

ChildView.cpp; (class CChildView : public CWnd， OnPaint() is here)
MainFrm.cpp; (class CMainFrame : public CFrameWnd)
ImagePro.cpp;(class CImageProApp : public CWinApp)

I also create CModalLessDlg using the dialog template, which is defined in ModalLessDlg.cpp and ModalLessDlg.h

In the  ModalLessDlg.cpp, I want to press the "Begin" button to send the "view" a message

you can realize it like this:

1. in ChildView.cpp

#include "ModalLessDlg.h"

void CChildView::OnModalLess() // show the modalless dialog
{

    CModalLessDlg* instance = new CModalLessDlg(this);
    instance->Create(CModalLessDlg::IDD,this);
    instance->PARENT = this;  
    instance->ShowWindow(1);

}

2. in ModalLessDlg.h

#include "ChildView.h"



public:

CChildView *PARENT;

3. in ChildView.h

public:

void     setDisplayText(CString src);

private:

CString   displayText;

4. in ChildView.cpp

CChildView::CChildView(): displayText(_T("zhiguang"))
{
}



void CChildView::setDisplayText(CString src)
{
    displayText=src;
    Invalidate(TRUE);
}



void CChildView::OnPaint() // show different results of image 
{


  CPaintDC dc(this); // device context for painting   
  CRect rect;
  GetClientRect (&rect);
  dc.DrawText (displayText, -1, &rect,DT_SINGLELINE | DT_CENTER | DT_VCENTER);

}

5. in ModalLessDlg.cpp

void CModalLessDlg::OnBegin() // 
{

  CString str_sent="OnBegin is running!";
  PARENT->setDisplayText(str_sent);
}

that is all.

it is using ChildView as one parameter to construct the modaless dialog, then you can easily use the ChildView and its members to change the client area.

SendMessage from modaless dialog to the view in SDI without document/view structure support

I create one SDI without document/view structure support, i name the project as ImagePro, then I can get the following files:

ChildView.cpp; (class CChildView : public CWnd， OnPaint() is here)
MainFrm.cpp; (class CMainFrame : public CFrameWnd)
ImagePro.cpp;(class CImageProApp : public CWinApp)

I also create CModalLessDlg using the dialog template, which is defined in ModalLessDlg.cpp and ModalLessDlg.h

In the  ModalLessDlg.cpp, I want to press the "Begin" button to send a text to the "view"

you can realize it like this:

1. in MainFrm.h, there is one protected variable:

protected: CChildView   m_wndView;

change it as public variable:

public: CChildView   m_wndView;

2. in ModalLessDlg.cpp:

#define WM_MY_DOSOME (WM_USER+1)

3. in ModalLessDlg.cpp:

void CModalLessDlg::OnBegin() // 
{
 // TODO: start
    
  CMainFrame   *pFrame=(CMainFrame*)(AfxGetApp()->m_pMainWnd); 
  char buf[]="I LOVE YOU";
 (pFrame->m_wndView).SendMessage(WM_MY_DOSOME,(WPARAM)buf,0);

}

4. In ChildView.h, add the following:

afx_msg LRESULT OnMyMessage(WPARAM wParam, LPARAM lParam);//before DECLARE_MESSAGE_MAP()

5. in ChildView.cpp, add the following:

BEGIN_MESSAGE_MAP(CChildView, CWnd)
 // ...
    ON_MESSAGE(WM_MY_DOSOME,OnMyMessage)  //only need to add this line

 //...
END_MESSAGE_MAP()

//then realize the response function:

LRESULT CChildView::OnMyMessage(WPARAM iParam1,LPARAM iParam2)
{
   
 char *buf=(char *)iParam1; 
 CString str;
 str.Format("%s",buf);
 MessageBox(str);
 return 0L;

}

the function is only a test to see whether you can receive the text or not, you can also show it in the client area of the view through Invalidate() again in this OnMyMessage function.

modaless dialog refresh the view in SDI without document/view structure support

I create one SDI without document/view structure support, i name the project as ImagePro, then I can get the following files:

ChildView.cpp; (class CChildView : public CWnd， OnPaint() is here)
MainFrm.cpp; (class CMainFrame : public CFrameWnd)
ImagePro.cpp;(class CImageProApp : public CWinApp)

I also create CModalLessDlg using the dialog template, which is defined in ModalLessDlg.cpp and ModalLessDlg.h

In the  ModalLessDlg.cpp, I want to press the "Begin" button to refresh the "view"

you can realize this like this:

1. in MainFrm.h, there is one protected variable:

protected: CChildView   m_wndView;

change it as public variable:

public: CChildView   m_wndView;

2. add one public variable for the ChildView.h:

public: int dlg_fresh;

initialize it in ChildView.cpp:

CChildView::CChildView()
{
 dlg_fresh=0;
}

3. in the response function of modalless dialog button, write like this:

//ModalLessDlg.cpp 




void CModalLessDlg::OnBegin() // stard the thread
{
 CMainFrame   *pFrame=(CMainFrame*)(AfxGetApp()->m_pMainWnd); // get the mainframe
 (pFrame->m_wndView).dlg_fresh=1; //set the flag
 (pFrame->m_wndView).Invalidate(); // refresh the view in the ChildView

}

4. in the ChildView.cpp, implement the OnPaint() like this:

//ChildView.cpp




void CChildView::OnPaint() 
{
 CPaintDC dc(this); // device context for painting
 CRect rect;
 GetClientRect (&rect);
 if(dlg_fresh==0)
 {
  dc.DrawText (_T ("it is not the modaless dlg refreshing!"), -1, &rect,DT_SINGLELINE | DT_CENTER | DT_VCENTER);  
 }
 else
 {
    dc.DrawText (_T ("it is the modaless dlg refreshing!"), -1, &rect,DT_SINGLELINE | DT_CENTER | DT_VCENTER);  
       dlg_fresh=0;
 }

}

kernel event object and thread in MFC

one button to create the kernel event object and worker thread, and the other button to signal the event and terminate the thread.

UINT __cdecl MyThreadProc(LPVOID lpParameter) // thread generation function
{
   CEvent* pEvent = (CEvent*)(lpParameter);
   VERIFY(pEvent != NULL);

   // Wait for the event to be signaled
   ::WaitForSingleObject(pEvent->m_hObject, INFINITE);

   // Terminate the thread
   ::AfxEndThread(0, FALSE); 
   return 0L;
}





void CModalLessDlg::OnBnClickedButton1() // start the thread
{
 // TODO: start
   pEvent = new CEvent(FALSE, FALSE);

   // Create a thread that will wait on the event
   //CWinThread* pThread;
   pThread = ::AfxBeginThread(&MyThreadProc, pEvent, 0, 0, CREATE_SUSPENDED, NULL);
   pThread->m_bAutoDelete = FALSE; 
   pThread->ResumeThread();
   
   // show the status of thread
   CString st1="Thread started";
   SetDlgItemText(IDC_EDIT1, st1);

}





void CModalLessDlg::OnBnClickedButton2() // setevent and terminate the thread
{


    pEvent->SetEvent();

   // Wait for the thread to consume the event and return
    ::WaitForSingleObject(pThread->m_hThread, INFINITE); 
    delete pThread;
    delete pEvent;

 // show the status of thread
    CString st2="Thread Terminated!";
    SetDlgItemText(IDC_EDIT1, st2);
}

edge detection based on sobel operator in MFC

void CChildView::OnEdge() // edge detection based on Sobel operators
{

 {
 int operx[3][3]={-1,0,1,-2,0,2,-1,0,1};//Sobel operators
 int opery[3][3]={1,2,1,0,0,0,-1,-2,-1};//Sobel operators
 int i,j;
 int p=0;
 int smx=0,smy=0;
 //CImage img_tem;
 //img_tem.Create(img_width,img_hight,24,0);
 for(i=0;i<img_hight;i++)
 {
  for(j=0;j<img_width;j++)
  {

   COLORREF cf = MyRGB(0,0,0);
   img_final.SetPixel(j,i,cf); //img_final is a member of this class, this is to reset it to zeros
         
  }
 }



 for(i=0;i<img_hight-2;i++)
 {
  for(j=0;j<img_width-2;j++)
  {

            smx=0;
   smy=0;
   for (int m=0; m<3; m++)
   {
      for (int n=0; n<3; n++)
      {   
      COLORREF rgb=img_new.GetPixel(j+n,i+m);
         int p=MyGetRValue(rgb);
                  smx=smx+p*operx[m][n];
                  smy=smy+p*opery[m][n];
      }
   }

   float G=smx*smx+smy*smy;
   int GD=(int)sqrt(G);
            
   if(GD>100) // threshold to determine whether it is an edge
   {
         COLORREF cf = MyRGB(255,255,255);
         img_final.SetPixel(j+1,i+1,cf);
 
   }

  }
 }
  
   Invalidate(); //call the OnPaint to show the result


}

basic image processing in MFC

Everybody knows that MFC+OpenCV is very convenient for image processing. But how about only using MFC? Most of the operation for image processing is based on bitmap in previous MFC version, say VC++6.0.

In VS2008, there is already CImage class, which could greatly facilitate the image processing.
 (1) get basic info of one image

  #include <atlimage.h>

  ...

  CImage img; //image object

  img.Load(FilePathName); // load a local image
  img_hight=img.GetHeight();//get hight
  img_width=img.GetWidth();//get width
  img_bits=img.GetBPP();//get bit counts for each pixel, say 8,24,32...

  COLORREF rgb=img.GetPixel(10,10);// get the pixel color

  



  CImage img_new; //another image object

  img_new.Create(img_width, img_hight, 24, 0);//create a image with the same size

  img.SetPixel(10,10,rgb);// set the pixel color

(2) get the r,g,b value from the COLORREF value:

#define MyGetRValue(rgb) ((BYTE)(rgb))
#define MyGetGValue(rgb) ((BYTE)(((WORD)(rgb)) >> 8))
#define MyGetBValue(rgb) ((BYTE)((rgb)>>16)) 



 COLORREF rgb=img.GetPixel(j,i);
 int r = MyGetRValue(rgb);

 int g=MyGetGValue(rgb);

 int b=MyGetBValue(rgb);

(3) convert the 24bit color image into gray image:

 #define MyRGB(r,g,b) ((COLORREF)(((BYTE)(r)|((WORD)((BYTE)(g))<<8))|(((DWORD)(BYTE)(b))<<16)))



 BYTE cl = (BYTE)(0.299 * r + 0.587 * g + 0.114 * b);
 int clr=(int)cl;
 COLORREF cf = MyRGB(clr,clr,clr);
  img_gray.SetPixel(j,i,cf);

 
 

read and write .ini file in MFC

Sometimes we need to configure the controls on the dialog based on the .ini and we may also need to save the final control values on the dialog to the same file in MFC, we can do that like this:

user.ini
[initialize]                   ;section name
m_listbox=0               ;key and int value       
name=zhiguang          ;key and string value


(1) example to initialize the controls on the dialog

BOOL CModalDlg::OnInitDialog() // read values from the .ini file
{
  
  CDialog::OnInitDialog();
  CString  iniFilePath;
  iniFilePath=".//ini//user.ini";
  // read ini for listbox
  int  list_pos;  
  list_pos=GetPrivateProfileInt("initialize","m_listbox",0,iniFilePath);
  m_listbox.AddString(_T("MIT"));
  m_listbox.AddString(_T("I.PAC"));
  m_listbox.AddString(_T("CASEM"));
  m_listbox.AddString(_T("Genric Power"));
  m_listbox.SetCurSel(list_pos);
  
  // read ini for edit
  CString  str_edit;  
  GetPrivateProfileString("initialize","IDC_EDIT1","",str_edit.GetBuffer(MAX_PATH),MAX_PATH,iniFilePath);
  SetDlgItemText(IDC_EDIT1, str_edit);

  // read ini for checkbox
  int  check_val;  
  check_val=GetPrivateProfileInt("initialize","m_check",0,iniFilePath);
  m_check.SetCheck(check_val);
  
  // read ini for radio
  int  radio_val;  
  radio_val=GetPrivateProfileInt("initialize","IDC_RADIO1",0,iniFilePath);
  if(radio_val==1)
  ((CButton *)GetDlgItem(IDC_RADIO1))->SetCheck(TRUE);
  else
  ((CButton *)GetDlgItem(IDC_RADIO2))->SetCheck(TRUE);

  return TRUE;
}

(2) example to update the control values on the dialog back to the same .ini file.

void CModalDlg::OnBnClickedOk() //save the values to the .ini file
{
 CString  iniFilePath;
    iniFilePath=".//ini//user.ini";

 //write ini for listbox
 int list_pos=m_listbox.GetCurSel();
 CString str_pos;
 str_pos.Format("%d",list_pos);
 WritePrivateProfileString("initialize","m_listbox",str_pos,iniFilePath);

 //write ini for editbox
 CString str_edit;
 GetDlgItemText(IDC_EDIT1, str_edit);
 WritePrivateProfileString("initialize","IDC_EDIT1",str_edit,iniFilePath);

   // read ini for checkbox
    int  check_val;  
    check_val=m_check.GetCheck();
 CString str_check;
 str_check.Format("%d",check_val);
 WritePrivateProfileString("initialize","m_check",str_check,iniFilePath);

   // read ini for radio button
    int  radio_val;  
    radio_val=((CButton *)GetDlgItem(IDC_RADIO1))->GetCheck();
 CString str_radio;
 str_radio.Format("%d",radio_val);
 WritePrivateProfileString("initialize","IDC_RADIO1",str_radio,iniFilePath);

 OnOK();
}

To read the file, we have  GetPrivateProfileString and GetPrivateProfileInt for String and Int values, to write the file , we only have WritePrivateProfileString, so we should convert the format of the value into CString first before writing

modal and modalless dialog in MFC

For both cases, you should create one dialog template, which could be done in the resource explorer

Modal Dialog
1. then create one class ModalDlg for modal dialog using this dialog template, and the CModalDlg.cpp and CModalDlg.h would be generated.

2. in the file where you want to show the modal dialog, first include the "ModalDlg.h", then in this file, you could use the following code :

void CChildView::OnModal() // show the modal dialog
{

 CModalDlg dlg;
 dlg.DoModal();
}

ModalLess Dialog

1. then create one class ModalLessDlg for modalless dialog using this dialog template, and the CModalLessDlg.cpp and CModalLessDlg.h would be generated.

2. in the file where you want to show the modalless dialog, first include the "ModalLessDlg.h", then in this file, you could use the following code :

void CChildView::OnModalLess() // show the modalless dialog
{
 CModalLessDlg *lsdlg=new CModalLessDlg;
 lsdlg->Create(IDD_DIALOG2,NULL);//IDD_DIALOG2 is the ID of dialog template you created
 lsdlg->ShowWindow(SW_SHOW);

 }

For both dialogs, if you want to initialize some parameters for all the controls on them, you can do that by overriding the  BOOL CModalDlg::OnInitDialog(){}  or  BOOL CModalLessDlg::OnInitDialog() {}

reserved keywords in mysql

today,  I created 3 mysql tables for my facebook app, to save user "status" "comment" "like" infomation.

I did not think too much before i created the 3 tables with name: status, comment, like respectively.
then, I found I could insert content into status and comment table while always failed the like table .

/*php code*/

mysql_query("INSERT INTO like (object_id,user_id,type) VALUES ('$comment_id', '$like_user_id','comment')");//error

I did know the 'like' is one mysql reserved keyword, however, i did not think that way at that very moment

then when I changed the table name from 'like' to 'likes', then it works.

/*php code*/

mysql_query("INSERT INTO likes (object_id,user_id,type) VALUES ('$comment_id', '$like_user_id','comment')");//correct

//or
mysql_query("INSERT INTO my_like (object_id,user_id,type) VALUES ('$comment_id', '$like_user_id','comment')");//correct

Here is the official mysql reserved keyword.
http://dev.mysql.com/doc/refman/5.0/en/reserved-words.html

however, I think it's safe to name your table or column as: my_xxx.

facebook fql format

pay attention to the "+" in the fql

 // 1.

 $access_token="XXX";

 $url0 = "https://graph.facebook.com/"

    . "fql?q=SELECT+message,time,status_id+FROM+status+WHERE+uid=10000+"

    . "AND+time>=1365481000+AND+time<=1365483292+LIMIT+0,1"

    . "&access_token=" . $access_token;



 $res0 = json_decode(file_get_contents($url0));






//2.



$url1 = "https://graph.facebook.com/"
                   . "fql?q=SELECT+fromid,time,text,object_id+FROM+comment+WHERE+object_id+"

                   ."IN+(SELECT+status_id+FROM+status+WHERE+uid=$uid)+"

                   ."AND+time>=$tm1+AND+time<=$tm2"
                   . "&access_token=" . $access_token;

$res1 = json_decode(file_get_contents($url1));

get realtime update of user infomation using facebook app

you can configure the framework according to the official document at:

http://developers.facebook.com/docs/reference/api/realtime/

if succeed, you can get the infomation like:

{ "object": "user", "entry":
    [    
         { "uid": 1335845740,      
           "changed_fields":
               [ "status" ],
           "time": 1365483292 }
   ]
}

It means the user(whose id is 1335845740 ) updated his status (most likely to post a new status).

what I want to say is, the 'time' 1365483292 you got is not exactly correct.

I use fql to crawl the status data and get the status created date is around 1365483292, but not exactly the same, say 1365483291. However, for some other times, they are  same.

So, if you got the update info as above, you should pay attention that "time" infomation.
if you still want to use this "time" info to crawl some latest data, you'd better use:

WHERE time BETWEEN $time-1 AND $time  // or something similar 

issue on inserting text or varchar values on mysql

when we insert text or varchar values into tables on mysql, we usually use the following format

mysql_query("INSERT INTO user (userid, name, gender) VALUES ('$user_id', '$user_name', '$user_gender'");

instead of

mysql_query("INSERT INTO user (userid, name, gender) VALUES ($user_id, $user_name, $user_gender");

yes, the difference is '$user_name' and  $user_name.

the advantage of former is that:
if your name is zhiguang cao (yes, there is one blank space between given name and surname )

when you insert  $user_name, it will be considered as two items: zhiguang  and cao respectively, and thus the result would not be correct generally.

while if you insert '$user_name', it will be considered as one item: 'zhiguang cao' would be one item instead of two.

I took me hours to find this problem although it seems not a big deal. thanks chenbo's help!