There is a comment on here, on a stackflow post that answers the question of how to achieve super fast median calculation on opencv. The question is here:
super fast median of matrix in opencv (as fast as matlab)
The problem is that the code is on C/C++ and I can't find a way to convert this to opencv for java or C#, since I'm trying to port this to OpenCV for Xamarin.Forms
Anyone knows how to conver this?
double medianMat(cv::Mat Input, int nVals) {
// COMPUTE HISTOGRAM OF SINGLE CHANNEL MATRIX
float range[] = {
0,
nVals
};
const float * histRange = {
range
};
bool uniform = true;
bool accumulate = false;
cv::Mat hist;
calcHist( & Input, 1, 0, cv::Mat(), hist, 1, & nVals, & histRange, uniform, accumulate);
// COMPUTE CUMULATIVE DISTRIBUTION FUNCTION (CDF)
cv::Mat cdf;
hist.copyTo(cdf);
for (int i = 1; i <= nVals - 1; i++) {
cdf.at < float > (i) += cdf.at < float > (i - 1);
}
cdf /= Input.total();
// COMPUTE MEDIAN
double medianVal;
for (int i = 0; i <= nVals - 1; i++) {
if (cdf.at < float > (i) >= 0.5) {
medianVal = i;
break;
}
}
return medianVal / nVals;
}
UPDATE:
As an example, this is my tried method on c# to calculate the mediam of a grey scale Mat. It fails, can you see what I did wrong?
private static int Median2(Mat Input)
{
// COMPUTE HISTOGRAM OF SINGLE CHANNEL MATRIX
MatOfInt mHistSize = new MatOfInt(256);
MatOfFloat mRanges = new MatOfFloat(0f, 256f);
MatOfInt channel = new MatOfInt(0); // only 1 channel for grey
Mat temp = new Mat();
Mat hist = new Mat();
Imgproc.CalcHist(Java.Util.Arrays.AsList(Input).Cast<Mat>().ToList(), channel, temp, hist, mHistSize, mRanges);
float[] cdf = new float[(int)hist.Total()];
hist.Get(0, 0, cdf);
for (int i = 1; i < 256; i++)
{
cdf[i] += cdf[i - 1];
cdf[i - 1] = cdf[i - 1] / Input.Total();
}
// COMPUTE CUMULATIVE DISTRIBUTION FUNCTION (CDF)
float total = (float)Input.Total();
// COMPUTE MEDIAN
double medianVal=0;
for (int i = 0; i < 256; i++)
{
if (cdf[i] >= 0.5) {
medianVal = i;
break;
}
}
return (int)(medianVal / 256);
}
Related
im struggeling for a long while now. Im trying to change the alpha channel on a specific pixel of a image. I want to convert an image to a 2D array with only alpha channels. I got this working in Android, but don't know how in UWP.
Android version
short[][] dotMapArray;
short[][] heatMapArray;
int[] scolorGradiant;
void init() {
Bitmap dotBitmap = loadImageFromResources(DOT);
Bitmap gradiantBitmap = loadImageFromResources(SPECTRUMPIC);
// initializing the 'dot' for the points on the heatmap
dotMapArray = new short[dotBitmap.getWidth()][dotBitmap.getHeight()];
print(true, "Converting Dot image to black image");
for (int x = 0; x < dotBitmap.getWidth(); x++) {
for (int y = 0; y < dotBitmap.getHeight(); y++) {
dotMapArray[x][y] = (short) (Color.alpha(dotBitmap.getPixel(x, y)));
if (dotMapArray[x][y] > 255)
dotMapArray[x][y] = 255;
}
}
// initializing the color gradiant
scolorGradiant = new int[gradiantBitmap.getHeight()];
int gradiant, pixelColor;
for (int y = 0; y < scolorGradiant.length; y++) {
gradiant = gradiantBitmap.getPixel(0, y); // color
pixelColor = Color.argb(90, Color.red(gradiant), Color.green(gradiant), Color.blue(gradiant));
scolorGradiant[y] = pixelColor;
}
}
How do I do this in UWP?
Can anybody help me?
Here's a loop that checks 1st channel (BGRA) if blue = 255 then all channels go to 0
var pixels = yourWriteableBitmap.PixelBuffer.ToArray();
for (int i = 0; i < pixels.Length; i += 4)
{
if (pixels[i] == 255)
{
pixels[i] = 0;
pixels[i + 1] = 0;
pixels[i + 2] = 0;
pixels[i + 3] = 0;
}
}
using (var stream = yourWriteableBitmap.PixelBuffer.AsStream())
{
await stream.WriteAsync(pixels, 0, pixels.Length);
}
I have a custom ImageTextButton in which I render the button to a FrameBuffer first and then draw with frameBuffer.getColorBufferTexture(). I don't really want to do this but I use a custom shader with this button that creates some visual effects and the only way I have been able to achieve it is with a FrameBuffer. I was surprised to find this actually works very smooth and fast though, the whole process takes 1-2ms on slow devices and having several instances doesn't cause any kind of framerate drop, so I am happy with this bit.
The issue I am having though is when I enable clipping on the ImageTextButton (with setClip(true)). The reason for this is the button can change in width, and I would like it to clip the text within the bounds of the button. If I disable the FrameBuffer and render normally, this part also works very well. If I combine the 2, it seems the clipping process gets confused and the result is either no text or very small parts of the text.
So here is the relevant code. I assumed it was because I set the FrameBuffer and SpriteBatch size/projection matrix just to deal with the active area (for efficiency) however if I don't modify any of this and use the same batch/projection matrix, so the FrameBuffer manages the whole screen, it is still the same result.
public void initFrameBuffer(){
xCache = (int) super.getX(); yCache = (int) super.getY();
widthCache = (int) super.getWidth(); heightCache = (int) super.getHeight();
frameBuffer = new FrameBuffer(Pixmap.Format.RGBA8888, widthCache, heightCache, false);
fboProjectionMatrix.setToOrtho2D(xCache, yCache+heightCache, widthCache, -heightCache);
this.fbBatch = new SpriteBatch();
this.fbBatch.setProjectionMatrix(fboProjectionMatrix);
this.frameBufferReady = true;
}
public void doFrameBuffer(Batch batch, float parentAlpha){
batch.end();
frameBuffer.begin();
fbBatch.begin();
Gdx.gl20.glClearColor(0f, 0.0f, 0.0f, 0.0f);
Gdx.gl20.glClear(GL20.GL_COLOR_BUFFER_BIT);
super.draw(fbBatch, parentAlpha);
fbBatch.end();
frameBuffer.end();
batch.begin();
}
public void drawFrameBufferObject(Batch batch, float parentAlpha){
batchColorCache = batch.getColor();
batch.setColor(1.0f, 1.0f, 1.0f, parentAlpha);
batch.draw(frameBuffer.getColorBufferTexture(), getX(), getY());
batch.setColor(batchColorCache);
}
#Override
public void draw(Batch batch, float parentAlpha) {
if (!this.frameBufferReady) initFrameBuffer();
doFrameBuffer(batch, parentAlpha);
drawFrameBufferObject(batch, parentAlpha);
}
Sorry for the long code, it's actually heavily trimmed down for the necessary parts..
Help hugely appreciated as always!
After much playing, the solution I have found is one that could probably be useful in other situations, and that is true clipping of the BitmapFontCache by vertex modification, no scissors involved! So if anyone would find this useful, the code is;
float xStart = ...start position of clip
float xEnd = ...end position of clip
//vertex offset numbers
int x_1 = 0, x_2 = 5, x2_1 = 10, x2_2 = 15;
int u_1 = 3, u_2 = 8, u2_1 = 13, u2_2 = 18;
for (int j = 0, n = pageCount; j < n; j++) {
int c = cache.getVertexCount(j);
int newIdx = 0;
if (c > 0) { // ignore if this texture has no glyphs
float[] vertices = cache.getVertices(j);
for(int i = 0; i < vertices.length; i+=20){
//if any of the vertices are outside the label, don't put them in the new cache
if(vertices[i+x2_1] > xStart && vertices[i+x_1] < xEnd){
for(int k = 0; k < 20; k++){
clippedVerts[j][newIdx+k] = vertices[i+k];
}
//case on major left glyph
if(vertices[i+x_1] < xStart){
float xDiff = vertices[i+x2_1]-xStart; //difference between right of glyph and clip
float xRatio = xDiff / (vertices[i+x2_1]-vertices[i+x_1]);
float uDiff = vertices[i+u2_1] - vertices[i+u_1];
float newU = vertices[i+u2_1] - uDiff*xRatio;
clippedVerts[j][newIdx+x_1] = xStart;
clippedVerts[j][newIdx+x_2] = xStart;
clippedVerts[j][newIdx+u_1] = newU;
clippedVerts[j][newIdx+u_2] = newU;
}
//case on major right glyph
if(vertices[i+x2_1] > xEnd){
float xDiff = xEnd-vertices[i+x_1]; //difference between left of glyph and clip
float xRatio = xDiff / (vertices[i+x2_1]-vertices[i+x_1]);
float uDiff = vertices[i+u2_1] - vertices[i+u_1];
float newU_2 = vertices[i+u_1] + uDiff*xRatio;
clippedVerts[j][newIdx+x2_1] = xEnd;
clippedVerts[j][newIdx+x2_2] = xEnd;
clippedVerts[j][newIdx+u2_1] = newU_2;
clippedVerts[j][newIdx+u2_2] = newU_2;
}
newIdx += 20;
}
}
}
clippedIdx[j] = newIdx;
}
for (int j = 0, n = pageCount; j < n; j++) {
int idx = clippedIdx[j];
if (idx > 0) { // ignore if this texture has no glyphs
float[] vertices = clippedVerts[j];
batch.draw(regions.get(j).getTexture(), vertices, 0, idx);
}
}
I am trying to draw Circle with texture on it which should be stretched on all vertices.
The problem is that the result i get looks like this:
http://s14.postimg.org/3wyb74469/image.png
I have tried to draw triangle fan as it need to be , first coordinates at 0,0,0
And rest as needed:
http://escience.anu.edu.au/lecture/cg/surfaceModeling/image/surfaceModeling015.png
Also here is the same question and I couldn't get answer for my problem from it :
OpenGL ES, add texture to circle
Loading circle vertices coordinates function:
private final int mVerticesDataSize = 3;
private final int mNumberOfVertices = 180;
private final int mBytesPerFloat = 4;
private float[] vertices;
private FloatBuffer mVerticesBuff;
public void loadCircleVerticesBuff(Context mActivityContext){
mVerticesBuff = ByteBuffer.allocateDirect(mNumberOfVertices * mVerticesDataSize * mBytesPerFloat).order(ByteOrder.nativeOrder()).asFloatBuffer();
vertices = new float[mNumberOfVertices * mVerticesDataSize];
float theta = 0;
for (int i = 0; i < (mNumberOfVertices * mVerticesDataSize); i += 3) {
vertices[i] = (float) (((float) 5*Math.cos(theta)));
vertices[i + 1] = (float) ((float) 5*Math.sin(theta));
vertices[i + 2] = 0;
theta += Math.PI / 90;
}
mVerticesBuff.put(vertices);
mVerticesBuff.position(0);
}
Loading circle texture coordinates function:
private final int mTextureCoordinateDataSize = 3;
public void loadCircleTextureBuff(){
mCircleTextureCoordinatesBuff = ByteBuffer.allocateDirect(mNumberOfVertices * mTextureCoordinateDataSize * mBytesPerFloat).order(ByteOrder.nativeOrder()).asFloatBuffer();
mCircleTextureCoordinatesBuff.put(vertices);
mCircleTextureCoordinatesBuff.position(0);
}
The opengl function used to draw is :
GLES20.glDrawArrays(GLES20.GL_TRIANGLE_FAN, 0, mNumberOfVertices);
So solution found (: 1 very very importent thing i missed up / or openGL missed up.
Texture coordinates can be only in range of 0 to 1 .
So here is the trick/solution :
//Build vertices :
vertices[i] = (float) (((float) raduis*Math.cos(theta)))+raduis;
vertices[i + 1] = (float) ((float) raduis*Math.sin(theta))+raduis;
// Build texture :
for (int i = 0; i < (mNumberOfVertices * mVerticesDataSize); i += 3) {
vertices[i] = (vertices[i])/(raduis*2);
vertices[i + 1] = (vertices[i+1])/(raduis*2);
vertices[i + 2] = 0;
}
And here is the res :
http://s2.postimg.org/tno4jr4y1/image.png
Dont forget to flip texture vertices as i forgot (:
i am trying to create gradient of two colors like Photoshop.
r,g,b of two colors is input and result will be the Mat of gradient. I tried it like for 5 hours at least and i could not find exact effect as of the Photoshop.
I tried to create my own formula (as i could not find any on the web), by changing RGB to HSV and then adding the difference of hue, with respect to the total number of rows, to each row of Mat with and also decreasing intensity to the center of image and then increasing it again. The code is self explanatory.
Additionally if anyone can tell me the exact formula for creating a gradient it will be really helpful.
here is how Photoshop gradient looks like
and this is what i get from my code
int r1, g1, b1, r2, g2, b2;
r1 = 255;
g1 = 0;
b1 = 0;
r2 = 0;
g2 = 255;
b2 = 0;
Mat input = imread("img.jpg");
Mat color1(input.size(), input.type());
Mat color2(input.size(), input.type());
vector<Mat> bgr1;
vector<Mat> bgr2;
split(color1, bgr1);
bgr1[0] = b1;
bgr1[1] = g1;
bgr1[2] = r1;
merge(bgr1, color1);
split(color2, bgr2);
bgr2[0] = b2;
bgr2[1] = g2;
bgr2[2] = r2;
merge(bgr2, color2);
vector<Mat> hls1;
vector<Mat> hls2;
cvtColor(color1, color1, CV_BGRA2BGR);
cvtColor(color1, color1, CV_BGR2HSV);
split(color1, hls1);
cvtColor(color2, color2, CV_BGRA2BGR);
cvtColor(color2, color2, CV_BGR2HSV);
split(color2, hls2);
double h1 = hls1[0].at<uchar>(0, 0);
double h2 = hls2[0].at<uchar>(0, 0);
double dif = (h2 - h1) / input.rows;
double h = h1;
double halfL = 255 / 2;
double halfR = input.rows / 2;
double ldif = halfL / halfR;
double l = 255;
bool isHalf = false;
for (int i = 0; i < input.rows; i++)
{
for (int j = 0; j < input.cols; j++)
{
hls1[0].at<uchar>(i, j) = h;
hls1[2].at<uchar>(i, j) = l;
}
if (isHalf == false){
l -= ldif;
}
else{
l += ldif;
}
if (i < input.rows * 0.40)
{
h += dif * 0.40;
}
else if (i < input.rows * 0.60)
{
h += dif * 3;
}
else
{
h += dif * 0.40;
}
if (i >= input.rows / 2)
{
isHalf = true;
}
}
merge(hls1, color1);
merge(hls2, color2);
cvtColor(color1, color1, CV_HSV2BGR);
cvtColor(color1, color1, CV_BGR2BGRA);
cvtColor(color2, color2, CV_HSV2BGR);
cvtColor(color2, color2, CV_BGR2BGRA);
namedWindow("Color1", cv::WINDOW_NORMAL);
resizeWindow("Color1", color1.size().width / 2, color1.size().height / 2);
imshow("Color1", color1);
waitKey(0);
destroyAllWindows();
system("pause");
I corrected my first code
It seems to be a really complex code for something that should be easier.
I would do something like that.
int taille = 500;
Mat image(taille,taille,CV_8UC3);
for(int y = 0; y < taille; y++){
Vec3b val;
val[0] = 0; val[1] = (y*255)/taille; val[2] = (taille-y)*255/taille;
for(int x = 0; x < taille; x++)
image.at<Vec3b>(y,x) = val;
}
On Micka's advice, I add a picture of the result with taille = 400;
I am doing histogram equalization on an image. I first get the RGB image and convert it to YUV. I run the histogram equalization algorithm on Y' of YUV and then convert back to RGB. Is it me, or does the image look weird? I am doing this correctly? this image is pretty bright, other images are a little red.
Here are the before/after images:
The algorithm (the commented values are values that I used previously for conversion. Both yield pretty much the same results) :
public static void createContrast(Bitmap src) {
int width = src.getWidth();
int height = src.getHeight();
Bitmap processedImage = Bitmap.createBitmap(width, height, src.getConfig());
int A = 0,R,G,B;
int pixel;
float[][] Y = new float[width][height];
float[][] U = new float[width][height];
float[][] V = new float [width][height];
int [] histogram = new int[256];
Arrays.fill(histogram, 0);
int [] cdf = new int[256];
Arrays.fill(cdf, 0);
float min = 257;
float max = 0;
for(int x = 0; x < width; ++x) {
for(int y = 0; y < height; ++y) {
pixel = src.getPixel(x, y);
//Log.i("TEST","("+x+","+y+")");
A = Color.alpha(pixel);
R = Color.red(pixel);
G = Color.green(pixel);
B = Color.blue(pixel);
/*Log.i("TESTEST","R: "+R);
Log.i("TESTEST","G: "+G);
Log.i("TESTEST","B: "+B);*/
// convert to YUV
/*Y[x][y] = 0.299f * R + 0.587f * G + 0.114f * B;
U[x][y] = 0.492f * (B-Y[x][y]);
V[x][y] = 0.877f * (R-Y[x][y]);*/
Y[x][y] = 0.299f * R + 0.587f * G + 0.114f * B;
U[x][y] = 0.565f * (B-Y[x][y]);
V[x][y] = 0.713f * (R-Y[x][y]);
// create a histogram
histogram[(int) Y[x][y]]+=1;
// get min and max values
if (Y[x][y] < min){
min = Y[x][y];
}
if (Y[x][y] > max){
max = Y[x][y];
}
}
}
cdf[0] = histogram[0];
for (int i=1;i<=255;i++){
cdf[i] = cdf[i-1] + histogram[i];
//Log.i("TESTEST","cdf of: "+i+" = "+cdf[i]);
}
float minCDF = cdf[(int)min];
float denominator = width*height - minCDF;
//Log.i("TEST","Histeq Histeq Histeq Histeq Histeq Histeq");
for(int x = 0; x < width; ++x) {
for(int y = 0; y < height; ++y) {
//Log.i("TEST","("+x+","+y+")");
pixel = src.getPixel(x, y);
A = Color.alpha(pixel);
Y[x][y] = ((cdf[ (int) Y[x][y]] - minCDF)/(denominator)) * 255;
/*R = minMaxCalc(Y[x][y] + 1.140f * V[x][y]);
G = minMaxCalc (Y[x][y] - 0.395f * U[x][y] - 0.581f * V[x][y]);
B = minMaxCalc (Y[x][y] + 2.032f * U[x][y]);*/
R = minMaxCalc(Y[x][y] + 1.140f * V[x][y]);
G = minMaxCalc (Y[x][y] - 0.344f * U[x][y] - 0.714f * V[x][y]);
B = minMaxCalc (Y[x][y] + 1.77f * U[x][y]);
//Log.i("TESTEST","A: "+A);
/*Log.i("TESTEST","R: "+R);
Log.i("TESTEST","G: "+G);
Log.i("TESTEST","B: "+B);*/
processedImage.setPixel(x, y, Color.argb(A, R, G, B));
}
}
}
My next step is to graph the histograms before and after. I just want to get an opinion here.
The question is a little bit old, but let me answer.
The reason is the way histogram equalization works. The algorithm tries to use all of the 0-255 range instead of given image's range.
So if you give it a dark image, it will change relatively brighter pixels to white colors. And relatively darker colors to black colors.
If you give it a bright image, for the same reason it will get darkened.