Is not issue of premute2 method? 
Posted: 11 January 2009 10:29 AM   [ Ignore ]
Rank
Joined  2007-12-16
Total Posts:  4
New Member

if(current[start].error[current[start].closest] > reject_distance_threshold)
break;
ids[start] = current[start].closest;

but I found that in front of ProcessResult method,it has already modifid the error and the closest correspondence.

for(i=0; i<cursize; i++)
{
// Bubble sort closest.
for(j=0; j<prevsize; j++)
{
for(k=0; k<prevsize-1-j; k++)
{
if(current.error[current.closest[k+1]] < current.error[current.closest[k]])
{
int tmp = current.closest[k]; // swap
current.closest[k] = current.closest[k+1];
current.closest[k+1] = tmp;

}
}
}
}

Profile
 
 
Posted: 11 January 2009 10:38 AM   [ Ignore ]   [ # 1 ]
Rank
Joined  2007-12-16
Total Posts:  4
New Member
Tony - 11 January 2009 10:29 AM

if(current[start].error[current[start].closest] > reject_distance_threshold)
break;
ids[start] = current[start].closest;

but I found that in front of ProcessResult method,it has already modifid the error and the closest correspondence.

for(i=0; i<cursize; i++)
{
// Bubble sort closest.
for(j=0; j<prevsize; j++)
{
for(k=0; k<prevsize-1-j; k++)
{
if(current.error[current.closest[k+1]] < current.error[current.closest[k]])
{
int tmp = current.closest[k]; // swap
current.closest[k] = current.closest[k+1];
current.closest[k+1] = tmp;
}
}
}
}

sorry! 
I am a careless.
the content of closest is error’s position.

Profile