How should I implement the Grassfire Algorithm in C++ - c++

So in my program, I generate a random grid using 2D Arrays where all indexes are initialized to 0. Now, a certain percentage of random indexes are filled with -1 which means that they are impassable/ act like a wall. The user also inputs a certain target index say (i,j) from where he starts and his goal is to reach index (0,0) by taking the shortest path possible.
To find the shortest path, I have to check for the neighbours of each cell, starting from the target location. If they have neighbours, I increment the neighbour value by 1. Refer to my figure for more details. I got the code on how to calculate the shortest path, but I'm stuck with this incrementation part. I tried writing a code but it doesn't seem to work. Any help would be appreciated:-
GRID is generated in the following way:
1 is the user input location, and the goal is to reach X i.e 0,0
-X 0 0 0 0 0 0 0 0 -1
-0 0 0 -1 -1 0 0 0 0 0
-0 0 0 0 -1 0 0 0 0 0
-0 0 0 0 0 0 0 0 0 -1
-0 0 0 0 0 0 0 1 0 0
Starting by incrementing
-X 0 0 0 0 0 0 0 0 -1
-0 0 0 -1 -1 0 0 0 0 0
-0 0 0 0 -1 3 3 3 3 3
-0 0 0 0 0 3 2 2 2 -1
-0 0 0 0 0 3 2 1 2 3
I have only showed it till 3, but it keeps on going until index 0,0 is reached.
void waveAlgorithm(int *array, int height, int width, int x, int y)
{
while (array != NULL)
{
// Assume that index 0 0 is never 1
if (currX == 0 && currY == 0){
break;
}
// Check South
int currX = x;
int currY = y + 1;
if (currX < width && currX > 0 && currY < height && currY >= 0)
{
if (*(array + currX * width + currY) == 0)
{
(*(array + currX * width + currY))++;
}
}
// Check North
currX = x;
currY = y - 1;
if (currX < width && currX > 0 && currY < height && currY >= 0)
{
if (*(array + currX * width + currY) != -1)
{
(*(array + currX * width + currY))++;
}
}
// Check West
currX = x - 1;
currY = y;
if (currX < width && currX > 0 && currY < height && currY >= 0)
{
if (*(array + currX * width + currY) != -1)
{
(*(array + currX * width + currY))++;
}
}
// Check East
currX = x + 1;
currY = y;
if (currX < width && currX > 0 && currY < height && currY >= 0)
{
if (*(array + currX * width + currY) != -1)
{
(*(array + currX * width + currY))++;
}
}
}
}
I am kinda stuck while implementing this program, especially for the the directions that are combinational i.e North East, South East, etc. I tried writing a recursive program but couldn't figure out how to increment the cells
waveAlgorithm(int *arr)
{
if(index is 0,0)
return;
waveAlgorithm(int[i+1][j]);
waveAlgorithm(int[i][j+1]);
waveAlgorithm(int[i-1][j]);
waveAlgorithm(int[i][j-1]);
}

Related

What can cause a Stack Overflow in this function?

I was writing a casual minesweeper, and wanted to realize a method to track an empty cells at the field, so I had wrote this algorothm:
//bigger array was taken to prevent out of range,when init mines and numbers
/*creating mines in 1 to FIELD_NUM range(0 -non-active field,1-active field)
* {
* 0 0 0 0 0 0
* 0 1 1 1 1 0
* 0 1 1 1 1 0
* 0 1 1 1 1 0
* 0 1 1 1 1 0
* 0 0 0 0 0 0
* }
*/
//view is a display vect,x and y are mouse interp. coord.
void MinerField::OpenCell(vector<vector<int>>& view, int x, int y)
{
if (gridLogic[x][y] == 9)
{
for (int i = 1; i <= FIELD_NUM; i++)
for (int j = 1; j <= FIELD_NUM; j++)
{
view[i][j] = gridLogic[i][j];
}
}
else
{
if (gridLogic[x][y] == 0)
OpenVoidCells(view, x, y);
else
view[x][y] = gridLogic[x][y];
}
}
And the second func,that is causing Stack-Overflow:
void MinerField::OpenVoidCells(vector<vector<int>>& view, int x, int y)
{
if (x >= (FIELD_NUM) || y >= (FIELD_NUM))//check out of range
return;
if (gridLogic[x][y] == 10 || gridLogic[x][y]==11 ||gridLogic[x][y]==-1)
return;
if ((gridLogic[x][y] <= 8) && (gridLogic[x][y] >= 1))
{
view[x][y] = gridLogic[x][y];
return;
}
view[x][y] = gridLogic[x][y];
OpenVoidCells(view,x + 1, y); //North;
OpenVoidCells(view,x - 1, y); //South
OpenVoidCells(view,x, y + 1); //East
OpenVoidCells(view, x, y - 1); //West
OpenVoidCells(view, x - 1, y - 1); //South-West
OpenVoidCells(view, x + 1, y + 1); //North-East
OpenVoidCells(view, x - 1, y + 1); //South-East
OpenVoidCells(view, x + 1, y - 1); //North-West
}
gridLogic vector is MinerField local and have the same size as view. Run fails with FIELD_NUM=10.
What can cause a stack overflow?
OpenVoidCells doesn't have anything to prevent visiting the same square over and over. It will go north, south, north, south, north, south ... forever, until you run out of stack. You need to keep track of visited squares and avoid re-checking them.

Problem with rotating an object with face to another one. [C++]

I want to rotate an object with the face side to the center of another one, but I have some problems with it: when I try to rotate an object to another one and it lies on X axis, it works properly [first two screenshots], but when I try to rotate it as on the screenshot, everything breaks down [second two screenshots].
Before1:
1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1
After1:
0 0 -1 0
-0 -1 0 0
1 0 0 0
0 0 0 1
Before2:
0 0 -1 0
-0 -1 0 0
1 0 0 0
0 0 0 1
After2:
0 0 -0.707107 0
0.5 -0.5 0 0
0.707107 -0.707107 0 0
0 0 0 1
Here's my code:
void ConcreteObject::faceObjectTo(ConcreteObject otherObject) {
Vector<double> temp = {0, 1, 0};
Vector<double> forward = otherObject.getCenter() - this->getCenter();
forward.normalize();
Vector<double> right = temp.cross(forward);
right.normalize();
Vector<double> up = forward.cross(right);
Matrix<double> newMatrix = this->getTransformMatrix().getCurrentState();
newMatrix(0, 0) = right[0];
newMatrix(0, 1) = right[1];
newMatrix(0, 2) = right[2];
newMatrix(1, 0) = up[0];
newMatrix(1, 1) = up[1];
newMatrix(1, 2) = up[2];
newMatrix(2, 0) = forward[0];
newMatrix(2, 1) = forward[1];
newMatrix(2, 2) = forward[2];
TransformMatrix newObjectMatrix(newMatrix);
this->setTransformMatrix(newObjectMatrix);
}
You need to normalize right, there's no reason for temp and forward to be orthogonal, hence even if they are unit vectors, their crossproduct need not be.

Swap two colors using color matrix

How can I swap two colors using a color matrix? For instance swapping red and blue is easy. The matrix would look like:
0 0 1 0 0
0 1 0 0 0
1 0 0 0 0
0 0 0 1 0
0 0 0 0 1
So how can I swap any two colors in general? For example, there is Color1 with R1, G1, B1 and Color2 with R2, G2, B2.
EDIT: By swap I mean Color1 will translate into color2 and color2 will translate into color1. Looks like I need a reflection transformation. How to calculate it?
GIMP reference removed. Sorry for confusion.
This appears to be the section of the color-exchange.c file in the GIMP source that cycles through all the pixels and if a pixel meets the chosen criteria(which can be a range of colors), swaps it with the chosen color:
for (y = y1; y < y2; y++)
{
gimp_pixel_rgn_get_row (&srcPR, src_row, x1, y, width);
for (x = 0; x < width; x++)
{
guchar pixel_red, pixel_green, pixel_blue;
guchar new_red, new_green, new_blue;
guint idx;
/* get current pixel-values */
pixel_red = src_row[x * bpp];
pixel_green = src_row[x * bpp + 1];
pixel_blue = src_row[x * bpp + 2];
idx = x * bpp;
/* want this pixel? */
if (pixel_red >= min_red &&
pixel_red <= max_red &&
pixel_green >= min_green &&
pixel_green <= max_green &&
pixel_blue >= min_blue &&
pixel_blue <= max_blue)
{
guchar red_delta, green_delta, blue_delta;
red_delta = pixel_red > from_red ?
pixel_red - from_red : from_red - pixel_red;
green_delta = pixel_green > from_green ?
pixel_green - from_green : from_green - pixel_green;
blue_delta = pixel_blue > from_blue ?
pixel_blue - from_blue : from_blue - pixel_blue;
new_red = CLAMP (to_red + red_delta, 0, 255);
new_green = CLAMP (to_green + green_delta, 0, 255);
new_blue = CLAMP (to_blue + blue_delta, 0, 255);
}
else
{
new_red = pixel_red;
new_green = pixel_green;
new_blue = pixel_blue;
}
/* fill buffer */
dest_row[idx + 0] = new_red;
dest_row[idx + 1] = new_green;
dest_row[idx + 2] = new_blue;
/* copy alpha-channel */
if (has_alpha)
dest_row[idx + 3] = src_row[x * bpp + 3];
}
/* store the dest */
gimp_pixel_rgn_set_row (&destPR, dest_row, x1, y, width);
/* and tell the user what we're doing */
if (!preview && (y % 10) == 0)
gimp_progress_update ((gdouble) y / (gdouble) height);
}
EDIT/ADDITION
Another way you could have transformed red to blue would be with this matrix:
1 0 0 0 0
0 1 0 0 0
0 0 1 0 0
0 0 0 1 0
-1 0 1 0 1
The only values that really matter are the bottom ones in this matrix.
This would be the same as saying subtract 255 from red, keep green the same, and then add 255 to blue. You could cut the alpha in half like so as well like so:
-1 0 1 -0.5 1
So (just like the gimp source) you just need to find the difference between your current color and your target color, for each channel, and then apply the difference. Instead of channel values from 0 to 255 you would use values from 0 to 1.
You could have changed it from red to green like so:
-1 1 0 0 1
See here for some good info:
http://msdn.microsoft.com/en-us/library/windows/desktop/ms533875%28v=vs.85%29.aspx
Good luck.
I solved it by creating a reflection matrix via D3DXMatrixReflect using a plane that's perpendicular to the vector AB and intersects the midpoint of the AB.
D3DXVECTOR3 AB( colorA.r-colorB.r, colorA.g-colorB.g, colorA.b-colorB.b );
D3DXPLANE plane( AB.x, AB.y, AB.z, -AB.x*midpoint.x-AB.y*midpoint.y-AB.z*midpoint.z );
D3DXMatrixReflect

Converting 1-bit bmp file to array in C/C++ [closed]

Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 2 years ago.
Improve this question
I'm looking to turn a 1-bit bmp file of variable height/width into a simple two-dimensional array with values of either 0 or 1. I don't have any experience with image editing in code and most libraries that I've found involve higher bit-depth than what I need. Any help regarding this would be great.
Here's the code to read a monochrome .bmp file
(See dmb's answer below for a small fix for odd-sized .bmps)
#include <stdio.h>
#include <string.h>
#include <malloc.h>
unsigned char *read_bmp(char *fname,int* _w, int* _h)
{
unsigned char head[54];
FILE *f = fopen(fname,"rb");
// BMP header is 54 bytes
fread(head, 1, 54, f);
int w = head[18] + ( ((int)head[19]) << 8) + ( ((int)head[20]) << 16) + ( ((int)head[21]) << 24);
int h = head[22] + ( ((int)head[23]) << 8) + ( ((int)head[24]) << 16) + ( ((int)head[25]) << 24);
// lines are aligned on 4-byte boundary
int lineSize = (w / 8 + (w / 8) % 4);
int fileSize = lineSize * h;
unsigned char *img = malloc(w * h), *data = malloc(fileSize);
// skip the header
fseek(f,54,SEEK_SET);
// skip palette - two rgb quads, 8 bytes
fseek(f, 8, SEEK_CUR);
// read data
fread(data,1,fileSize,f);
// decode bits
int i, j, k, rev_j;
for(j = 0, rev_j = h - 1; j < h ; j++, rev_j--) {
for(i = 0 ; i < w / 8; i++) {
int fpos = j * lineSize + i, pos = rev_j * w + i * 8;
for(k = 0 ; k < 8 ; k++)
img[pos + (7 - k)] = (data[fpos] >> k ) & 1;
}
}
free(data);
*_w = w; *_h = h;
return img;
}
int main()
{
int w, h, i, j;
unsigned char* img = read_bmp("test1.bmp", &w, &h);
for(j = 0 ; j < h ; j++)
{
for(i = 0 ; i < w ; i++)
printf("%c ", img[j * w + i] ? '0' : '1' );
printf("\n");
}
return 0;
}
It is plain C, so no pointer casting - beware while using it in C++.
The biggest problem is that the lines in .bmp files are 4-byte aligned which matters a lot with single-bit images. So we calculate the line size as "width / 8 + (width / 8) % 4". Each byte contains 8 pixels, not one, so we use the k-based loop.
I hope the other code is obvious - much has been told about .bmp header and pallete data (8 bytes which we skip).
Expected output:
0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0
0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0
0 0 0 0 0 0 1 1 1 1 0 0 1 1 0 0
0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0
0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0
0 0 0 1 0 0 1 1 1 1 0 0 0 0 0 0
0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0
0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0
0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0
0 0 0 0 0 0 1 1 1 1 0 0 1 0 0 0
0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0
0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0
0 0 0 0 0 1 1 1 1 1 0 0 0 0 1 0
0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0
0 0 0 1 0 1 1 1 1 1 0 0 0 0 0 0
0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0
I tried the solution of Viktor Lapyov on a 20x20 test image:
But with his code, I get this output (slightly reformatted but you can see the problem):
The last 4 pixels are not read. The problem is here. (The last partial byte in a row is ignored.)
// decode bits
int i, j, k, rev_j;
for(j = 0, rev_j = h - 1; j < h ; j++, rev_j--) {
for(i = 0 ; i < w / 8; i++) {
int fpos = j * lineSize + i, pos = rev_j * w + i * 8;
for(k = 0 ; k < 8 ; k++)
img[pos + (7 - k)] = (data[fpos] >> k ) & 1;
}
}
I rewrote the inner loop like this:
// decode bits
int i, byte_ctr, j, rev_j;
for(j = 0, rev_j = h - 1; j < h ; j++, rev_j--) {
for( i = 0; i < w; i++) {
byte_ctr = i / 8;
unsigned char data_byte = data[j * lineSize + byte_ctr];
int pos = rev_j * w + i;
unsigned char mask = 0x80 >> i % 8;
img[pos] = (data_byte & mask ) ? 1 : 0;
}
}
and all is well:
The following c code works with monochrome bitmaps of any size. I'll assume you've got your bitmap in a buffer with heights and width initialized from file. So
// allocate mem for global buffer
if (!(img = malloc(h * w)) )
return(0);
int i = 0, k, j, scanline;
// calc the scanline. Monochrome images are
// padded with 0 at every line end. This
// makes them divisible by 4.
scanline = ( w + (w % 8) ) >> 3;
// account for the paddings
if (scanline % 4)
scanline += (4 - scanline % 4);
// loop and set the img values
for (i = 0, k = h - 1; i < h; i++)
for (j = 0; j < w; j++) {
img[j+i*w] = (buffer[(j>>3)+k*scanline])
& (0x80 >> (j % 8));
}
Hope this help's. To convert it to 2D is now a trivial matter: But if u get lost here is the math to convert 1D array to 2D suppose r & c are row and column and w is the width then:
. c + r * w = r, c
If you got further remarks hit me back, am out!!!
Lets think of a1x7 monochrome bitmap i.e. This is a bitmap of a straight line with 7 pixels wide. To store this image on a Windows OS; since 7 is not evenly divisible by 4 it's going to pad in it an extra 3 bytes.
So the biSizeImage of the BITMAPINFOHEADER structure will show a total of 4 bytes. Nonetheless the biHeight and biWidth members will correctly state the true bitmap dimensions.
The above code will fail because 7 / 8 = 0 (by rounding off as with all c compilers do). Hence loop "i" will not execute so will "k".
That means the vector "img" now contains garbage values that do not correspond to the pixels contained in " data" i.e. the result is incorrect.
And by inductive reasoning if it does not satisfy the base case then chances are it wont do much good for general cases.

Image downscaling algorithm

Could you help me find the right algorithm for image resizing? I have an image of a number. The maximum size is 200x200, I need to get an image with size 15x15 or even less. The image is monochrome (black and white) and the result should be the same. That's the info about my task.
I've already tried one algorithm, here it is
// xscale, yscale - decrease/increase rate
for (int f = 0; f<=49; f++)
{
for (int g = 0; g<=49; g++)//49+1 - final size
{
xpos = (int)f * xscale;
ypos = (int)g * yscale;
picture3[f][g]=picture4[xpos][ypos];
}
}
But it won't work with the decrease of an image, which is my prior target.
Could you help me find an algorithm, which could solve that problem (quality mustn't be perfect, the speed doesn't even matter). Some information about it would be perfect too considering the fact I'm a newbie. Of course, a short piece of c/c++ code (or a library) will be perfect too.
Edit:
I've found an algorithm. Will it be suitable for compressing from 200 to 20?
The general approach is to filter the input to generate a smaller size, and threshold to convert to monochrome. The easiest filter to implement is a simple average, and it often produces OK results. The Sinc filter is theoretically the best but it's impractical to implement and has ringing artifacts which are often undesirable. Many other filters are available, such as Lanczos or Tent (which is the generalized form of Bilinear).
Here's a version of an average filter combined with thresholding. Assuming picture4 is the input with pixel values of 0 or 1, and the output is picture3 in the same format. I also assumed that x is the least significant dimension which is opposite to the usual mathematical notation, and opposite to the coordinates in your question.
int thumbwidth = 15;
int thumbheight = 15;
double xscale = (thumbwidth+0.0) / width;
double yscale = (thumbheight+0.0) / height;
double threshold = 0.5 / (xscale * yscale);
double yend = 0.0;
for (int f = 0; f < thumbheight; f++) // y on output
{
double ystart = yend;
yend = (f + 1) / yscale;
if (yend >= height) yend = height - 0.000001;
double xend = 0.0;
for (int g = 0; g < thumbwidth; g++) // x on output
{
double xstart = xend;
xend = (g + 1) / xscale;
if (xend >= width) xend = width - 0.000001;
double sum = 0.0;
for (int y = (int)ystart; y <= (int)yend; ++y)
{
double yportion = 1.0;
if (y == (int)ystart) yportion -= ystart - y;
if (y == (int)yend) yportion -= y+1 - yend;
for (int x = (int)xstart; x <= (int)xend; ++x)
{
double xportion = 1.0;
if (x == (int)xstart) xportion -= xstart - x;
if (x == (int)xend) xportion -= x+1 - xend;
sum += picture4[y][x] * yportion * xportion;
}
}
picture3[f][g] = (sum > threshold) ? 1 : 0;
}
}
I've now tested this code. Here's the input 200x200 image, followed by a nearest-neighbor reduction to 15x15 (done in Paint Shop Pro), followed by the results of this code. I'll leave you to decide which is more faithful to the original; the difference would be much more obvious if the original had some fine detail.
To properly downscale an image, you should divide your image up into square blocks of pixels and then use something like Bilinear Interpolation in order to find the right color of the pixel that should replace the NxN block of pixels you're doing the interpolation on.
Since I'm not so good at the math involved, I'm not going to try give you an example of how the code would like. Sorry :(
Since you're fine with using a library, you could look into the imagemagick C++ bindings.
You could also output the image in a simple format like a pbm, and then call the imagemagick command to resize it:
system("convert input.pbm -resize 10x10 -compress none output.pbm");
Sample output file (note: you don't need to use a new line for each row):
P1
20 20
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 0 0 0 0 0
0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 0 0 0 0 0
0 0 0 0 0 0 0 1 1 0 0 0 0 1 1 0 0 0 0 0
0 0 0 0 0 1 1 1 1 0 0 0 0 1 1 1 0 0 0 0
0 0 0 0 0 1 1 1 1 0 0 0 0 1 1 1 0 0 0 0
0 0 0 0 0 1 1 1 1 0 0 0 0 1 1 1 0 0 0 0
0 0 0 0 0 1 1 1 1 0 0 0 0 1 1 1 0 0 0 0
0 0 0 0 0 0 0 1 1 0 0 0 0 1 1 1 0 0 0 0
0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 0 0
0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
The output file:
P1
10 10
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 1 1 0 1 1 0
0 0 0 0 1 0 0 1 1 0 0 0 0 0 1 0 0 1 1 0 0 0 0 0 1 1 0 1 1 0 0 0 0 0 0 1 1 1 1
1 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0
I've found an implementation of a bilinear interpolaton. C code.
Assuming that:
a - a primary array (which we need to stretch/compress) pointer.
oldw - primary width
oldh - primary height
b - a secondary array (which we get after compressing/stretching) pointer
neww - secondary width
newh - seconday height
#include <stdio.h>
#include <math.h>
#include <sys/types.h>
void resample(void *a, void *b, int oldw, int oldh, int neww, int newh)
{
int i;
int j;
int l;
int c;
float t;
float u;
float tmp;
float d1, d2, d3, d4;
u_int p1, p2, p3, p4; /* nearby pixels */
u_char red, green, blue;
for (i = 0; i < newh; i++) {
for (j = 0; j < neww; j++) {
tmp = (float) (i) / (float) (newh - 1) * (oldh - 1);
l = (int) floor(tmp);
if (l < 0) {
l = 0;
} else {
if (l >= oldh - 1) {
l = oldh - 2;
}
}
u = tmp - l;
tmp = (float) (j) / (float) (neww - 1) * (oldw - 1);
c = (int) floor(tmp);
if (c < 0) {
c = 0;
} else {
if (c >= oldw - 1) {
c = oldw - 2;
}
}
t = tmp - c;
/* coefficients */
d1 = (1 - t) * (1 - u);
d2 = t * (1 - u);
d3 = t * u;
d4 = (1 - t) * u;
/* nearby pixels: a[i][j] */
p1 = *((u_int*)a + (l * oldw) + c);
p2 = *((u_int*)a + (l * oldw) + c + 1);
p3 = *((u_int*)a + ((l + 1)* oldw) + c + 1);
p4 = *((u_int*)a + ((l + 1)* oldw) + c);
/* color components */
blue = (u_char)p1 * d1 + (u_char)p2 * d2 + (u_char)p3 * d3 + (u_char)p4 * d4;
green = (u_char)(p1 >> 8) * d1 + (u_char)(p2 >> 8) * d2 + (u_char)(p3 >> 8) * d3 + (u_char)(p4 >> 8) * d4;
red = (u_char)(p1 >> 16) * d1 + (u_char)(p2 >> 16) * d2 + (u_char)(p3 >> 16) * d3 + (u_char)(p4 >> 16) * d4;
/* new pixel R G B */
*((u_int*)b + (i * neww) + j) = (red << 16) | (green << 8) | (blue);
}
}
}
Hope it will be useful for other users. But nevertheless I still doubth whether it will work in my situation (when not stratching, but compressing an array). Any ideas?
I think, you need Interpolation. There are a lot of algorithms, for example you can use Bilinear interpolation
If you use Win32, then StretchBlt function possibly help.
The StretchBlt function copies a bitmap from a source rectangle into a destination rectangle, stretching or compressing the bitmap to fit the dimensions of the destination rectangle, if necessary. The system stretches or compresses the bitmap according to the stretching mode currently set in the destination device context.
One approach to downsizing a 200x200 image to, say 100x100, would be to take every 2nd pixel along each row and column. I'll leave you to roll your own code for downsizing to a size which is not a divisor of the original size. And I provide no warranty as to the suitability of this approach for your problem.