spelling changes and change to kernel size

This commit is contained in:
kimo-s 2023-11-15 13:59:51 -05:00
parent b6d07d2a29
commit 6d286a7837
4 changed files with 64 additions and 53 deletions

View File

@ -9,7 +9,7 @@
* Example licensed under an unmodified zlib/libpng license, which is an OSI-certified,
* BSD-like license that allows static linking with closed source software
*
* Copyright (c) 2015-2023 Ramon Santamaria (@raysan5)
* Copyright (c) 2015-2023 Karim Salem (@kimo-s)
*
********************************************************************************************/
@ -25,8 +25,8 @@ int main(void)
Image image = LoadImage("resources/cat.png"); // Loaded in CPU memory (RAM)
const int screenWidth = image.width*4;
const int screenHeight = image.height;
const int screenWidth = 800;
const int screenHeight = 450;
InitWindow(screenWidth, screenHeight, "raylib [textures] example - image convolution");
@ -43,18 +43,20 @@ int main(void)
0.0, -1.0, 0.0};
Image catSharpend = ImageCopy(image);
ImageKernelConvolution(&catSharpend, sharpenkernel, 3);
ImageKernelConvolution(&catSharpend, sharpenkernel, 16);
Image catSobel = ImageCopy(image);
ImageKernelConvolution(&catSobel, sobelkernel, 3);
ImageKernelConvolution(&catSobel, sobelkernel, 16);
Image catGaussian = ImageCopy(image);
for(int i = 0; i < 6; i++){
ImageKernelConvolution(&catGaussian, gaussiankernel, 3);
ImageKernelConvolution(&catGaussian, gaussiankernel, 16);
}
ImageCrop(&image, (Rectangle){ 0, 0, (float)200, (float)450 });
ImageCrop(&catGaussian, (Rectangle){ 0, 0, (float)200, (float)450 });
ImageCrop(&catSobel, (Rectangle){ 0, 0, (float)200, (float)450 });
ImageCrop(&catSharpend, (Rectangle){ 0, 0, (float)200, (float)450 });
Texture2D texture = LoadTextureFromImage(image); // Image converted to texture, GPU memory (VRAM)
Texture2D catSharpendTexture = LoadTextureFromImage(catSharpend);
Texture2D catSobelTexture = LoadTextureFromImage(catSobel);
@ -82,9 +84,9 @@ int main(void)
ClearBackground(RAYWHITE);
DrawTexture(catSharpendTexture, 0, 0, WHITE);
DrawTexture(catSobelTexture, texture.width, 0, WHITE);
DrawTexture(catGaussianTexture, texture.width*2, 0, WHITE);
DrawTexture(texture, texture.width*3, 0, WHITE);
DrawTexture(catSobelTexture, 200, 0, WHITE);
DrawTexture(catGaussianTexture, 400, 0, WHITE);
DrawTexture(texture, 600, 0, WHITE);
EndDrawing();
//----------------------------------------------------------------------------------

View File

@ -339,7 +339,7 @@ RLAPI void ImageAlphaClear(Image *image, Color color, float threshold);
RLAPI void ImageAlphaMask(Image *image, Image alphaMask); // Apply alpha mask to image
RLAPI void ImageAlphaPremultiply(Image *image); // Premultiply alpha channel
RLAPI void ImageBlurGaussian(Image *image, int blurSize); // Apply Gaussian blur using a box blur approximation
RLAPI void ImageKernelConvolution(Image *image, float* karnel, int karenlWidth); // Apply Custom Square image convolution kernel
RLAPI void ImageKernelConvolution(Image *image, float* kernel, int kernelSize); // Apply Custom Square image convolution kernel
RLAPI void ImageResize(Image *image, int newWidth, int newHeight); // Resize image (Bicubic scaling algorithm)
RLAPI void ImageResizeNN(Image *image, int newWidth,int newHeight); // Resize image (Nearest-Neighbor scaling algorithm)
RLAPI void ImageResizeCanvas(Image *image, int newWidth, int newHeight, int offsetX, int offsetY, Color fill); // Resize canvas and fill with color

View File

@ -1329,7 +1329,7 @@ RLAPI void ImageAlphaClear(Image *image, Color color, float threshold);
RLAPI void ImageAlphaMask(Image *image, Image alphaMask); // Apply alpha mask to image
RLAPI void ImageAlphaPremultiply(Image *image); // Premultiply alpha channel
RLAPI void ImageBlurGaussian(Image *image, int blurSize); // Apply Gaussian blur using a box blur approximation
RLAPI void ImageKernelConvolution(Image *image, float* karnel, int karenlWidth); // Apply Custom Square image convolution kernel
RLAPI void ImageKernelConvolution(Image *image, float* kernel, int kernelSize); // Apply Custom Square image convolution kernel
RLAPI void ImageResize(Image *image, int newWidth, int newHeight); // Resize image (Bicubic scaling algorithm)
RLAPI void ImageResizeNN(Image *image, int newWidth,int newHeight); // Resize image (Nearest-Neighbor scaling algorithm)
RLAPI void ImageResizeCanvas(Image *image, int newWidth, int newHeight, int offsetX, int offsetY, Color fill); // Resize canvas and fill with color

View File

@ -2083,90 +2083,100 @@ void ImageBlurGaussian(Image *image, int blurSize) {
}
// The kernel matrix is assumed to be square. Only supply the width of the kernel.
void ImageKernelConvolution(Image *image, float* karnel, int karenlWidth){
void ImageKernelConvolution(Image *image, float* kernel, int kernelSize){
if ((image->data == NULL) || (image->width == 0) || (image->height == 0) || karnel == NULL) return;
if ((image->data == NULL) || (image->width == 0) || (image->height == 0) || kernel == NULL) return;
ImageAlphaPremultiply(image);
int kernelWidth = (int)sqrtf((float)kernelSize);
if (kernelWidth*kernelWidth != kernelSize)
{
TRACELOG(LOG_WARNING, "IMAGE: Convolution kernel must be square to be applied");
return;
}
Color *pixels = LoadImageColors(*image);
Vector4 *imageCopy1 = RL_MALLOC((image->height)*(image->width)*sizeof(Vector4));
Vector4 *imageCopy2 = RL_MALLOC((image->height)*(image->width)*sizeof(Vector4));
Vector4 *temp = RL_MALLOC(karenlWidth*karenlWidth*sizeof(Vector4));
Vector4 *temp = RL_MALLOC(kernelSize*sizeof(Vector4));
float normKernel = 0.0f;
for(int i = 0; i < karenlWidth * karenlWidth; i++){
for(int i = 0; i < kernelSize; i++){
temp[i].x = 0.0f;
temp[i].y = 0.0f;
temp[i].z = 0.0f;
temp[i].w = 0.0f;
normKernel += karnel[i];
normKernel += kernel[i];
}
if(normKernel != 0.0f){
for(int i = 0; i < karenlWidth * karenlWidth; i++){
karnel[i] /= normKernel;
for(int i = 0; i < kernelSize; i++){
kernel[i] /= normKernel;
}
}
float rRes = 0.0f;
float gRes = 0.0f;
float bRes = 0.0f;
float aRes = 0.0f;
for (int i = 0; i < (image->height)*(image->width); i++) {
imageCopy1[i].x = ((float)pixels[i].r)/255.0f;
imageCopy1[i].y = ((float)pixels[i].g)/255.0f;
imageCopy1[i].z = ((float)pixels[i].b)/255.0f;
imageCopy1[i].w = ((float)pixels[i].a)/255.0f;
}
int startRange, endRange;
if(karenlWidth % 2 == 0){
startRange = -karenlWidth/2;
endRange = karenlWidth/2;
if(kernelWidth % 2 == 0){
startRange = -kernelWidth/2;
endRange = kernelWidth/2;
} else {
startRange = -karenlWidth/2;
endRange = karenlWidth/2+1;
startRange = -kernelWidth/2;
endRange = kernelWidth/2+1;
}
for(int x = 0; x < image->height; x++) {
for(int y = 0; y < image->width; y++) {
for(int xk = startRange; xk < endRange; xk++){
for(int yk = startRange; yk < endRange; yk++){
int xkabs = xk + karenlWidth/2;
int ykabs = yk + karenlWidth/2;
int xkabs = xk + kernelWidth/2;
int ykabs = yk + kernelWidth/2;
size_t imgindex = image->width * (x+xk) + (y+yk);
if(imgindex < 0 || imgindex >= image->width * image->height){
temp[karenlWidth * xkabs + ykabs].x = 0.0f;
temp[karenlWidth * xkabs + ykabs].y = 0.0f;
temp[karenlWidth * xkabs + ykabs].z = 0.0f;
temp[karenlWidth * xkabs + ykabs].w = 0.0f;
temp[kernelWidth * xkabs + ykabs].x = 0.0f;
temp[kernelWidth * xkabs + ykabs].y = 0.0f;
temp[kernelWidth * xkabs + ykabs].z = 0.0f;
temp[kernelWidth * xkabs + ykabs].w = 0.0f;
} else {
temp[karenlWidth * xkabs + ykabs].x = imageCopy1[imgindex].x * karnel[karenlWidth * xkabs + ykabs];
temp[karenlWidth * xkabs + ykabs].y = imageCopy1[imgindex].y * karnel[karenlWidth * xkabs + ykabs];
temp[karenlWidth * xkabs + ykabs].z = imageCopy1[imgindex].z * karnel[karenlWidth * xkabs + ykabs];
temp[karenlWidth * xkabs + ykabs].w = imageCopy1[imgindex].w * karnel[karenlWidth * xkabs + ykabs];
temp[kernelWidth * xkabs + ykabs].x = ((float)pixels[imgindex].r)/255.0f * kernel[kernelWidth * xkabs + ykabs];
temp[kernelWidth * xkabs + ykabs].y = ((float)pixels[imgindex].g)/255.0f * kernel[kernelWidth * xkabs + ykabs];
temp[kernelWidth * xkabs + ykabs].z = ((float)pixels[imgindex].b)/255.0f * kernel[kernelWidth * xkabs + ykabs];
temp[kernelWidth * xkabs + ykabs].w = ((float)pixels[imgindex].a)/255.0f * kernel[kernelWidth * xkabs + ykabs];
}
}
}
for(int i = 0; i < karenlWidth * karenlWidth; i++){
for(int i = 0; i < kernelSize; i++){
rRes += temp[i].x;
gRes += temp[i].y;
bRes += temp[i].z;
aRes += temp[i].w;
}
STBIR_CLAMP(rRes, 0.0f, 1.0f);
STBIR_CLAMP(gRes, 0.0f, 1.0f);
STBIR_CLAMP(bRes, 0.0f, 1.0f);
STBIR_CLAMP(aRes, 0.0f, 1.0f);
if(rRes < 0.0f){
rRes = 0.0f;
}
if(gRes < 0.0f){
gRes = 0.0f;
}
if(bRes < 0.0f){
bRes = 0.0f;
}
if(rRes > 1.0f){
rRes = 1.0f;
}
if(gRes > 1.0f){
gRes = 1.0f;
}
if(bRes > 1.0f){
bRes = 1.0f;
}
imageCopy2[image->width * (x) + (y)].x = rRes;
imageCopy2[image->width * (x) + (y)].y = gRes;
@ -2178,7 +2188,7 @@ void ImageKernelConvolution(Image *image, float* karnel, int karenlWidth){
bRes = 0.0f;
aRes = 0.0f;
for(int i = 0; i < karenlWidth * karenlWidth; i++){
for(int i = 0; i < kernelSize; i++){
temp[i].x = 0.0f;
temp[i].y = 0.0f;
temp[i].z = 0.0f;
@ -2199,7 +2209,6 @@ void ImageKernelConvolution(Image *image, float* karnel, int karenlWidth){
int format = image->format;
RL_FREE(image->data);
RL_FREE(imageCopy1);
RL_FREE(imageCopy2);
RL_FREE(temp);