package ij.plugin.filter;
import ij.*;
import ij.process.*;
import ij.gui.*;
import ij.measure.*;
import ij.plugin.ContrastEnhancer;
import java.awt.*;
import java.util.*;
/**
This class implements the Process/FFT/Bandpass Filter command. It is based on
Joachim Walter's FFT Filter plugin at "http://imagej.nih.gov/ij/plugins/fft-filter.html".
2001/10/29: First Version (JW)
2003/02/06: 1st bugfix (works in macros/plugins, works on stacks, overwrites image(=>filter)) (JW)
2003/07/03: integrated into ImageJ, added "Display Filter" option (WSR)
2007/03/26: 2nd bugfix (Fixed incorrect calculation of filter from structure sizes, which caused
the real structure sizes to be wrong by a factor of 0.75 to 1.5 depending on the image size.)
*/
public class FFTFilter implements PlugInFilter, Measurements {
private ImagePlus imp;
private String arg;
private static int filterIndex = 1;
private FHT fht;
private int slice;
private int stackSize = 1;
private static double filterLargeDia = 40.0;
private static double filterSmallDia = 3.0;
private static int choiceIndex = 0;
private static String[] choices = {"None","Horizontal","Vertical"};
private static String choiceDia = choices[0];
private static double toleranceDia = 5.0;
private static boolean doScalingDia = true;
private static boolean saturateDia = true;
private static boolean displayFilter;
private static boolean processStack;
public int setup(String arg, ImagePlus imp) {
this.arg = arg;
this.imp = imp;
if (imp==null)
{IJ.noImage(); return DONE;}
stackSize = imp.getStackSize();
fht = (FHT)imp.getProperty("FHT");
if (fht!=null) {
IJ.error("FFT Filter", "Spatial domain image required");
return DONE;
}
if (!showBandpassDialog(imp))
return DONE;
else
return processStack?DOES_ALL+DOES_STACKS+PARALLELIZE_STACKS:DOES_ALL;
}
public void run(ImageProcessor ip) {
slice++;
filter(ip);
}
void filter(ImageProcessor ip) {
ImageProcessor ip2 = ip;
if (ip2 instanceof ColorProcessor) {
showStatus("Extracting brightness");
ip2 = ((ColorProcessor)ip2).getBrightness();
}
Rectangle roiRect = ip2.getRoi();
int maxN = Math.max(roiRect.width, roiRect.height);
double sharpness = (100.0 - toleranceDia) / 100.0;
boolean doScaling = doScalingDia;
boolean saturate = saturateDia;
IJ.showProgress(1,20);
/* tile mirrored image to power of 2 size
first determine smallest power 2 >= 1.5 * image width/height
factor of 1.5 to avoid wrap-around effects of Fourier Trafo */
int i=2;
while(i<1.5 * maxN) i *= 2;
// Calculate the inverse of the 1/e frequencies for large and small structures.
double filterLarge = 2.0*filterLargeDia / (double)i;
double filterSmall = 2.0*filterSmallDia / (double)i;
// fit image into power of 2 size
Rectangle fitRect = new Rectangle();
fitRect.x = (int) Math.round( (i - roiRect.width) / 2.0 );
fitRect.y = (int) Math.round( (i - roiRect.height) / 2.0 );
fitRect.width = roiRect.width;
fitRect.height = roiRect.height;
// put image (ROI) into power 2 size image
// mirroring to avoid wrap around effects
showStatus("Pad to "+i+"x"+i);
ip2 = tileMirror(ip2, i, i, fitRect.x, fitRect.y);
IJ.showProgress(2,20);
// transform forward
showStatus(i+"x"+i+" forward transform");
FHT fht = new FHT(ip2);
fht.setShowProgress(false);
fht.transform();
IJ.showProgress(9,20);
//new ImagePlus("after fht",ip2.crop()).show();
// filter out large and small structures
showStatus("Filter in frequency domain");
filterLargeSmall(fht, filterLarge, filterSmall, choiceIndex, sharpness);
//new ImagePlus("filter",ip2.crop()).show();
IJ.showProgress(11,20);
// transform backward
showStatus("Inverse transform");
fht.inverseTransform();
IJ.showProgress(19,20);
//new ImagePlus("after inverse",ip2).show();
// crop to original size and do scaling if selected
showStatus("Crop and convert to original type");
fht.setRoi(fitRect);
ip2 = fht.crop();
if (doScaling) {
ImagePlus imp2 = new ImagePlus(imp.getTitle()+"-filtered", ip2);
new ContrastEnhancer().stretchHistogram(imp2, saturate?1.0:0.0);
ip2 = imp2.getProcessor();
}
// convert back to original data type
int bitDepth = imp.getBitDepth();
switch (bitDepth) {
case 8: ip2 = ip2.convertToByte(doScaling); break;
case 16: ip2 = ip2.convertToShort(doScaling); break;
case 24:
ip.snapshot();
showStatus("Setting brightness");
((ColorProcessor)ip).setBrightness((FloatProcessor)ip2);
break;
case 32: break;
}
// copy filtered image back into original image
if (bitDepth!=24) {
ip.snapshot();
ip.copyBits(ip2, roiRect.x, roiRect.y, Blitter.COPY);
}
ip.resetMinAndMax();
IJ.showProgress(20,20);
}
void showStatus(String msg) {
if (stackSize>1 && processStack)
IJ.showStatus("FFT Filter: "+slice+"/"+stackSize);
else
IJ.showStatus(msg);
}
/** Puts ImageProcessor (ROI) into a new ImageProcessor of size width x height y at position (x,y).
The image is mirrored around its edges to avoid wrap around effects of the FFT. */
public ImageProcessor tileMirror(ImageProcessor ip, int width, int height, int x, int y) {
if (IJ.debugMode) IJ.log("FFT.tileMirror: "+width+"x"+height+" "+ip);
if (x < 0 || x > (width -1) || y < 0 || y > (height -1)) {
IJ.error("Image to be tiled is out of bounds.");
return null;
}
ImageProcessor ipout = ip.createProcessor(width, height);
ImageProcessor ip2 = ip.crop();
int w2 = ip2.getWidth();
int h2 = ip2.getHeight();
//how many times does ip2 fit into ipout?
int i1 = (int) Math.ceil(x / (double) w2);
int i2 = (int) Math.ceil( (width - x) / (double) w2);
int j1 = (int) Math.ceil(y / (double) h2);
int j2 = (int) Math.ceil( (height - y) / (double) h2);
//tile
if ( (i1%2) > 0.5)
ip2.flipHorizontal();
if ( (j1%2) > 0.5)
ip2.flipVertical();
for (int i=-i1; i<i2; i += 2) {
for (int j=-j1; j<j2; j += 2) {
ipout.insert(ip2, x-i*w2, y-j*h2);
}
}
ip2.flipHorizontal();
for (int i=-i1+1; i<i2; i += 2) {
for (int j=-j1; j<j2; j += 2) {
ipout.insert(ip2, x-i*w2, y-j*h2);
}
}
ip2.flipVertical();
for (int i=-i1+1; i<i2; i += 2) {
for (int j=-j1+1; j<j2; j += 2) {
ipout.insert(ip2, x-i*w2, y-j*h2);
}
}
ip2.flipHorizontal();
for (int i=-i1; i<i2; i += 2) {
for (int j=-j1+1; j<j2; j += 2) {
ipout.insert(ip2, x-i*w2, y-j*h2);
}
}
return ipout;
}
/*
filterLarge: down to which size are large structures suppressed?
filterSmall: up to which size are small structures suppressed?
filterLarge and filterSmall are given as fraction of the image size
in the original (untransformed) image.
stripesHorVert: filter out: 0) nothing more 1) horizontal 2) vertical stripes
(i.e. frequencies with x=0 / y=0)
scaleStripes: width of the stripe filter, same unit as filterLarge
*/
void filterLargeSmall(ImageProcessor ip, double filterLarge, double filterSmall, int stripesHorVert, double scaleStripes) {
int maxN = ip.getWidth();
float[] fht = (float[])ip.getPixels();
float[] filter = new float[maxN*maxN];
for (int i=0; i<maxN*maxN; i++)
filter[i]=1f;
int row;
int backrow;
float rowFactLarge;
float rowFactSmall;
int col;
int backcol;
float factor;
float colFactLarge;
float colFactSmall;
float factStripes;
// calculate factor in exponent of Gaussian from filterLarge / filterSmall
double scaleLarge = filterLarge*filterLarge;
double scaleSmall = filterSmall*filterSmall;
scaleStripes = scaleStripes*scaleStripes;
//float FactStripes;
// loop over rows
for (int j=1; j<maxN/2; j++) {
row = j * maxN;
backrow = (maxN-j)*maxN;
rowFactLarge = (float) Math.exp(-(j*j) * scaleLarge);
rowFactSmall = (float) Math.exp(-(j*j) * scaleSmall);
// loop over columns
for (col=1; col<maxN/2; col++){
backcol = maxN-col;
colFactLarge = (float) Math.exp(- (col*col) * scaleLarge);
colFactSmall = (float) Math.exp(- (col*col) * scaleSmall);
factor = (1 - rowFactLarge*colFactLarge) * rowFactSmall*colFactSmall;
switch (stripesHorVert) {
case 1: factor *= (1 - (float) Math.exp(- (col*col) * scaleStripes)); break;// hor stripes
case 2: factor *= (1 - (float) Math.exp(- (j*j) * scaleStripes)); // vert stripes
}
fht[col+row] *= factor;
fht[col+backrow] *= factor;
fht[backcol+row] *= factor;
fht[backcol+backrow] *= factor;
filter[col+row] *= factor;
filter[col+backrow] *= factor;
filter[backcol+row] *= factor;
filter[backcol+backrow] *= factor;
}
}
//process meeting points (maxN/2,0) , (0,maxN/2), and (maxN/2,maxN/2)
int rowmid = maxN * (maxN/2);
rowFactLarge = (float) Math.exp(- (maxN/2)*(maxN/2) * scaleLarge);
rowFactSmall = (float) Math.exp(- (maxN/2)*(maxN/2) * scaleSmall);
factStripes = (float) Math.exp(- (maxN/2)*(maxN/2) * scaleStripes);
fht[maxN/2] *= (1 - rowFactLarge) * rowFactSmall; // (maxN/2,0)
fht[rowmid] *= (1 - rowFactLarge) * rowFactSmall; // (0,maxN/2)
fht[maxN/2 + rowmid] *= (1 - rowFactLarge*rowFactLarge) * rowFactSmall*rowFactSmall; // (maxN/2,maxN/2)
filter[maxN/2] *= (1 - rowFactLarge) * rowFactSmall; // (maxN/2,0)
filter[rowmid] *= (1 - rowFactLarge) * rowFactSmall; // (0,maxN/2)
filter[maxN/2 + rowmid] *= (1 - rowFactLarge*rowFactLarge) * rowFactSmall*rowFactSmall; // (maxN/2,maxN/2)
switch (stripesHorVert) {
case 1: fht[maxN/2] *= (1 - factStripes);
fht[rowmid] = 0;
fht[maxN/2 + rowmid] *= (1 - factStripes);
filter[maxN/2] *= (1 - factStripes);
filter[rowmid] = 0;
filter[maxN/2 + rowmid] *= (1 - factStripes);
break; // hor stripes
case 2: fht[maxN/2] = 0;
fht[rowmid] *= (1 - factStripes);
fht[maxN/2 + rowmid] *= (1 - factStripes);
filter[maxN/2] = 0;
filter[rowmid] *= (1 - factStripes);
filter[maxN/2 + rowmid] *= (1 - factStripes);
break; // vert stripes
}
//loop along row 0 and maxN/2
rowFactLarge = (float) Math.exp(- (maxN/2)*(maxN/2) * scaleLarge);
rowFactSmall = (float) Math.exp(- (maxN/2)*(maxN/2) * scaleSmall);
for (col=1; col<maxN/2; col++){
backcol = maxN-col;
colFactLarge = (float) Math.exp(- (col*col) * scaleLarge);
colFactSmall = (float) Math.exp(- (col*col) * scaleSmall);
switch (stripesHorVert) {
case 0:
fht[col] *= (1 - colFactLarge) * colFactSmall;
fht[backcol] *= (1 - colFactLarge) * colFactSmall;
fht[col+rowmid] *= (1 - colFactLarge*rowFactLarge) * colFactSmall*rowFactSmall;
fht[backcol+rowmid] *= (1 - colFactLarge*rowFactLarge) * colFactSmall*rowFactSmall;
filter[col] *= (1 - colFactLarge) * colFactSmall;
filter[backcol] *= (1 - colFactLarge) * colFactSmall;
filter[col+rowmid] *= (1 - colFactLarge*rowFactLarge) * colFactSmall*rowFactSmall;
filter[backcol+rowmid] *= (1 - colFactLarge*rowFactLarge) * colFactSmall*rowFactSmall;
break;
case 1:
factStripes = (float) Math.exp(- (col*col) * scaleStripes);
fht[col] *= (1 - colFactLarge) * colFactSmall * (1 - factStripes);
fht[backcol] *= (1 - colFactLarge) * colFactSmall * (1 - factStripes);
fht[col+rowmid] *= (1 - colFactLarge*rowFactLarge) * colFactSmall*rowFactSmall * (1 - factStripes);
fht[backcol+rowmid] *= (1 - colFactLarge*rowFactLarge) * colFactSmall*rowFactSmall * (1 - factStripes);
filter[col] *= (1 - colFactLarge) * colFactSmall * (1 - factStripes);
filter[backcol] *= (1 - colFactLarge) * colFactSmall * (1 - factStripes);
filter[col+rowmid] *= (1 - colFactLarge*rowFactLarge) * colFactSmall*rowFactSmall * (1 - factStripes);
filter[backcol+rowmid] *= (1 - colFactLarge*rowFactLarge) * colFactSmall*rowFactSmall * (1 - factStripes);
break;
case 2:
factStripes = (float) Math.exp(- (maxN/2)*(maxN/2) * scaleStripes);
fht[col] = 0;
fht[backcol] = 0;
fht[col+rowmid] *= (1 - colFactLarge*rowFactLarge) * colFactSmall*rowFactSmall * (1 - factStripes);
fht[backcol+rowmid] *= (1 - colFactLarge*rowFactLarge) * colFactSmall*rowFactSmall * (1 - factStripes);
filter[col] = 0;
filter[backcol] = 0;
filter[col+rowmid] *= (1 - colFactLarge*rowFactLarge) * colFactSmall*rowFactSmall * (1 - factStripes);
filter[backcol+rowmid] *= (1 - colFactLarge*rowFactLarge) * colFactSmall*rowFactSmall * (1 - factStripes);
}
}
// loop along column 0 and maxN/2
colFactLarge = (float) Math.exp(- (maxN/2)*(maxN/2) * scaleLarge);
colFactSmall = (float) Math.exp(- (maxN/2)*(maxN/2) * scaleSmall);
for (int j=1; j<maxN/2; j++) {
row = j * maxN;
backrow = (maxN-j)*maxN;
rowFactLarge = (float) Math.exp(- (j*j) * scaleLarge);
rowFactSmall = (float) Math.exp(- (j*j) * scaleSmall);
switch (stripesHorVert) {
case 0:
fht[row] *= (1 - rowFactLarge) * rowFactSmall;
fht[backrow] *= (1 - rowFactLarge) * rowFactSmall;
fht[row+maxN/2] *= (1 - rowFactLarge*colFactLarge) * rowFactSmall*colFactSmall;
fht[backrow+maxN/2] *= (1 - rowFactLarge*colFactLarge) * rowFactSmall*colFactSmall;
filter[row] *= (1 - rowFactLarge) * rowFactSmall;
filter[backrow] *= (1 - rowFactLarge) * rowFactSmall;
filter[row+maxN/2] *= (1 - rowFactLarge*colFactLarge) * rowFactSmall*colFactSmall;
filter[backrow+maxN/2] *= (1 - rowFactLarge*colFactLarge) * rowFactSmall*colFactSmall;
break;
case 1:
factStripes = (float) Math.exp(- (maxN/2)*(maxN/2) * scaleStripes);
fht[row] = 0;
fht[backrow] = 0;
fht[row+maxN/2] *= (1 - rowFactLarge*colFactLarge) * rowFactSmall*colFactSmall * (1 - factStripes);
fht[backrow+maxN/2] *= (1 - rowFactLarge*colFactLarge) * rowFactSmall*colFactSmall * (1 - factStripes);
filter[row] = 0;
filter[backrow] = 0;
filter[row+maxN/2] *= (1 - rowFactLarge*colFactLarge) * rowFactSmall*colFactSmall * (1 - factStripes);
filter[backrow+maxN/2] *= (1 - rowFactLarge*colFactLarge) * rowFactSmall*colFactSmall * (1 - factStripes);
break;
case 2:
factStripes = (float) Math.exp(- (j*j) * scaleStripes);
fht[row] *= (1 - rowFactLarge) * rowFactSmall * (1 - factStripes);
fht[backrow] *= (1 - rowFactLarge) * rowFactSmall * (1 - factStripes);
fht[row+maxN/2] *= (1 - rowFactLarge*colFactLarge) * rowFactSmall*colFactSmall * (1 - factStripes);
fht[backrow+maxN/2] *= (1 - rowFactLarge*colFactLarge) * rowFactSmall*colFactSmall * (1 - factStripes);
filter[row] *= (1 - rowFactLarge) * rowFactSmall * (1 - factStripes);
filter[backrow] *= (1 - rowFactLarge) * rowFactSmall * (1 - factStripes);
filter[row+maxN/2] *= (1 - rowFactLarge*colFactLarge) * rowFactSmall*colFactSmall * (1 - factStripes);
filter[backrow+maxN/2] *= (1 - rowFactLarge*colFactLarge) * rowFactSmall*colFactSmall * (1 - factStripes);
}
}
if (displayFilter && slice==1) {
FHT f = new FHT(new FloatProcessor(maxN, maxN, filter, null));
f.swapQuadrants();
new ImagePlus("Filter", f).show();
}
}
boolean showBandpassDialog(ImagePlus imp) {
GenericDialog gd = new GenericDialog("FFT Bandpass Filter");
gd.addNumericField("Filter_large structures down to", filterLargeDia, 0, 4, "pixels");
gd.addNumericField("Filter_small structures up to", filterSmallDia, 0, 4, "pixels");
gd.addChoice("Suppress stripes:", choices, choiceDia);
gd.addNumericField("Tolerance of direction:", toleranceDia, 0, 2, "%");
gd.addCheckbox("Autoscale after filtering", doScalingDia);
gd.addCheckbox("Saturate image when autoscaling", saturateDia);
gd.addCheckbox("Display filter", displayFilter);
if (stackSize>1)
gd.addCheckbox("Process entire stack", processStack);
gd.addHelp(IJ.URL+"/docs/menus/process.html#fft-bandpass");
gd.showDialog();
if(gd.wasCanceled())
return false;
if(gd.invalidNumber()) {
IJ.error("Error", "Invalid input number");
return false;
}
filterLargeDia = gd.getNextNumber();
filterSmallDia = gd.getNextNumber();
choiceIndex = gd.getNextChoiceIndex();
choiceDia = choices[choiceIndex];
toleranceDia = gd.getNextNumber();
doScalingDia = gd.getNextBoolean();
saturateDia = gd.getNextBoolean();
displayFilter = gd.getNextBoolean();
if (stackSize>1)
processStack = gd.getNextBoolean();
return true;
}
}