In This Topic ▼

L_TADAnisotropicDiffusion

Summary

An iterative filter that performs Tensor-guided Anisotropic Diffusion (TAD) in order to reduce noise while preserving an image's edges.

Syntax

#include "Ltimgcor.h"

L_LTIMGCOR_API L_INT L_TADAnisotropicDiffusion(pBitmap, nIterations, nLambda, nKappa, nFlags)

Parameters

pBITMAPHANDLE pBitmap

Pointer to the bitmap handle.

L_INT nIterations

Number of iterations the filter will perform. The default value is 10.

L_INT nLambda

An integer that represents the integration constant, which controls the speed of diffusion. Typically set it to the maximum (25). The default value is 14.

L_INT nKappa

Gradient modulus threshold that controls conduction. The default value is 30.
If Kappa is low, small intensity gradients are able to block conduction and hence diffusion across the step edges. A large value reduces the influence of intensity gradients on conduction.

L_UINT nFlags

The type of the conduction to use. The following values are possible:TAD_EXPON_FLUX = 0x00000001 Use exponential flux, which favors high contrast edges over low contrast ones.TAD_QUAD_FLUX = 0x00000002 Use quadratic flux, which favors wide regions over smaller ones. This is the default value.

Returns

Value	Meaning
SUCCESS	The function was successful.
< 1	An error occurred. Refer to Return Codes.

Comments

Use the L_TADAnisotropicDiffusion to reduce noise in images (more specifically, medical images).

LEADTOOLS offers three functions that implement anisotropic diffusion:

L_AnisotropicDiffusionBitmap: L_AnisotropicDiffusion is a slower but more accurate version of L_SRADAnisotropDiffusion.
L_SRADAnisotropicDiffusion: In comparison with conventional anisotropic diffusion, the uniform areas resulting from application of the L_SRADAnisotropicDiffusion function exhibit more smoothing and edges and features are preserved better.
L_TADAnisotropicDiffusion: L_TADAnisotropicDiffusion employs an edge-seeking diffusion coefficient.

All three reduce noise and preserve edges and perform better than the L_MedianFilterBitmap or L_GaussianFilterBitmap functions. With the anisotropic functions, areas within regions are smoothed, but the edges are not affected.

The number of iterations controls the number of times the filter will be applied to an image. The lower the number of iterations, the faster the filter will perform. Higher iterations typically mean clearer results.

This function can only process the entire image. It does not support regions.

This function supports 8- and 16-bit grayscale images and 24- and 32-bit colored images.

This function supports signed/unsigned images.

Required DLLs and Libraries

LTIMGCOR
For a listing of the exact DLLs and Libraries needed, based on the toolkit version, refer to Files To Be Included With Your Application.

Platforms

Win32, x64.

Example

This example loads a bitmap and applies TADAnisotropicDiffusion.

L_INT TADAnisotropicDiffusionExample(L_VOID) 
{ 
   L_INT nRet; 
   BITMAPHANDLE LeadBitmap;   /* Bitmap handle to hold the loaded image. */ 
 
   /* Load the bitmap, keeping the bits per pixel of the file */ 
   nRet = L_LoadBitmap(MAKE_IMAGE_PATH(TEXT("DICOM\\IMAGE3.dcm")), &LeadBitmap, sizeof(BITMAPHANDLE), 0, ORDER_BGR, NULL, NULL); 
   if (nRet != SUCCESS) 
      return nRet; 
 
   /* Apply TADAnisotropicDiffusion filter */ 
   nRet = L_TADAnisotropicDiffusion(&LeadBitmap, 10, 14, 30, TAD_QUAD_FLUX); 
 
   if (nRet == SUCCESS) 
      nRet = L_SaveBitmap(MAKE_IMAGE_PATH(TEXT("Result.BMP")), &LeadBitmap, FILE_BMP, 24, 0, NULL); 
 
   //free bitmap 
   if (LeadBitmap.Flags.Allocated) 
      L_FreeBitmap(&LeadBitmap); 
 
   return nRet; 
}