ImageJ plugin IJ_JTransforms

Requires ImageJ 1.53k or higher, ImageJAddins.jar, JTransforms-3.1-with-dependencies.jar.

• Forward FFTs
• Inverse FFTs
• FFT Operations
• Discrete Transforms
• Plot Transforms
• Plot Operations

Compute the Fourier, Sine, Cosine, and Hartley, forward and inverse tranforms of images, image Stacks, 3D images, and plots.

Compute Fourier correlation, convolution and deconvolution

Easy to use: Clone this project and add your own plugins using the IJJT_fft wrapper class to hide the conversions between ImageJ data formats and JTransforms sequenced format.
Examples:

	//instantiate the wrapper class and custom return class
	IJJT_fft ijft = new IJJT_fft();
	ComplexArr cpa = ijft.new ComplexArr();	
	
	//2D fft of current slice of an image stack
	Object sliceData  =  imp.getStack().getProcessor(curSlice).getPixels();
	cpa= ijft.realForward_2D(sliceData,width,height,centered,magPhase);								
	reImp.getProcessor().setPixels(cpa.re);
	imImp.getProcessor().setPixels(cpa.im);

	//2D convolution of the current slice of a target stack
	//with the current slice of a kernel stack
	int targSlice = targImp.getCurrentSlice();
	int kernSlice = kernImp.getCurrentSlice();
	Object oTarg = targImp.getStack().getProcessor(targSlice).getPixels();
	Object oKern = kernImp.getStack().getProcessor(kernSlice).getPixels();
	cpa = ijft.fourierOperation_2D("convolve", oTarg, oKern, targWidth, targHeight);
	corrImp = IJ.createImage(newTitle,targWidth, targHeight, 1, 32);
	corrImp.getProcessor().setPixels(cpa.re);

	//3D FFT of a stack
	Object[] oVolData = imp.getStack().getImageArray();									
	cpa = ijft.realForward_3D(oVolData, width, height, depth, centered, magPhase);			
	reImp.getStack().setVoxels(0, 0, 0, width, height, depth, cpa.re);
	imImp.getStack().setVoxels(0, 0, 0, width, height, depth, cpa.im);

Support: 8, 16, and 32 bit gray scale images. Inputs are converted to double precision. Outputs are converted to 32 bit.

Limitations:

• Data dimensions must be even values. Not a JTransforms limitation, restricting methods to even dimension values significantly simplifed this project
• Deconvolution may require advanced techniques for noisy data

Forward FFT

• Compute centered FFT - Places the DC term at the center of the transform image. Default is DC term at the (0,0) coordinate
• Show as Magnitude and Phase - Displays FFT as magnitude and phase. Default is real and imaginary.
• FFT - Selections are "Current Slice 2D", "All Slices 2D", and "All Slices 3D"
• Output is a new image(s) with the transform result.

Inverse FFT

• Real or Magnitude Image - can be an ImageJ image or Stack
• Imaginary or Phase Image - can be an ImageJ image or Stack
• FFT - Selections are "Current Slice 2D", "All Slices 2D", and "All Slices 3D". Select the correct one for the input images.
• Input is Centered FFT - Select the correct one for the input images.
• Input is Magnitude and Phase -Select the correct one for the input images.
• Output is a new image(s) with the transform result.

FFT Operations

• Target Image/Stack - can be an ImageJ image or Stack, the kernel will be applied to this image
• Kernel Image/Stack - can be an ImageJ image or Stack. Apply this image to the target
• Target and Kernel can be the same image, for example, autocorrelation, useful in texture analysis
• FFT Operations- Selections are "correlate", "convolve", and "deconvolve".
• Target and Kernel - Formats for the respective input images. Selections are:
  • Slice - Slice 2D - Use the current target and kernel slices.
  • Stack - Slice 2D - Apply the current kernel to every slice in a target stack.
  • Stack - Stack 2D - Apply the kernel to the target slice by slice.
  • Stack - Stack 3D - Apply the 3D kernel to the 3D target.
• Output is a new image(s) with the operation result.

Discrete Transforms

• Data - Selections are "Current Slice 2D", "All Slices 2D", and "All Slices 3D"
• Transform - Selections are "Cosine", "Sine", and "Hartley".
• Direction - foward or inverse.
• Image sizes may be odd values
• Output is a new image(s) with the transform result.

Plot Transforms

• Plot - Select the plot that has the data to transform.
• Real Data - Select the data from the selected plot.
• Imaginary Data - Only visible for inverse Fourier.
• Transform - Selections are "Fourier", "Cosine", "Sine", and "Hartley"
• Direction - Foward or Inverse.
• Compute Centered Transform - Only visible for Fourier and Hartley Transforms
• Show as Magnitude and Phase - Only visible for Fourier Transforms
• Output is a new plot with the operation result.

Plot Operations

• Target Plot - Select the plot that has the data to operate on.
• Target Data - Select the data from the target plot.
• Kernel Plot - Select the plot that has the data to apply to the target.
• kernel Data - Select the data to apply from the kernel plot.
• Operation - Selections are "correlate", "convolve", and "deconvolve"
• Output is a new plot with the operation result.