Scribd
Upload
50 views4 pages

Threshold and Wavelet

The document discusses different thresholding techniques used in digital image processing for image segmentation. It describes global, local, and adaptive thresholding. Adaptive thresholding is based on the difference in pixel intensities between regions and is best for medical image denoising. The paper uses an adaptive thresholding technique based on wavelet shrinkage and the Matlab software to denoise medical images like MRI and CT scans. Wavelet shrinkage removes noise while preserving important features by concentrating image features in large wavelet coefficients and setting small coefficients likely containing noise to zero.

Uploaded by

Eati Iswarya Lakshmi
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
50 views4 pages

Threshold and Wavelet

The document discusses different thresholding techniques used in digital image processing for image segmentation. It describes global, local, and adaptive thresholding. Adaptive thresholding is based on the difference in pixel intensities between regions and is best for medical image denoising. The paper uses an adaptive thresholding technique based on wavelet shrinkage and the Matlab software to denoise medical images like MRI and CT scans. Wavelet shrinkage removes noise while preserving important features by concentrating image features in large wavelet coefficients and setting small coefficients likely containing noise to zero.

Uploaded by

Eati Iswarya Lakshmi
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 4

THRESHOLD TECHNIQUES

According to digital image processing, thresholding is known for segmenting the


images .it is used to create a binary image which consists of black and white pixels which
have two intensity values. Threshold is a fixed value which draws a boundary line between
two set of data.

FIG-1 THRESHOLDING TECHNIQUE


The thresholding is one of the important techniques in image segmentation in the field of
image processing that is used to isolate the objects by converting grey-scale images.

The thresholding is calculated by using an iterative method .It is non-linear


operation. The threshold value is estimated by clustering method of the histogram
1) Firstly separates the pixels into two clusters

2) Find the mean of it and then

3) Square the difference between the means and

4) Multiply the by the number of pixels in one cluster times the number in the other.

And the equation for that calculation mentioned above is as follows-

σ2 Between(T) = nB(T)nO(T)[µB(T) − µO(T)]2

The types of threshold techniques in field of image processing are-

1. GLOBAL THRESHOLDING
2. LOCAL THRESHOLDING
3. ADAPTIVE THRESHOLDING

GLOBAL THRESHOLDING-
GLOBAL THRESHOLDING is a technique in which it consists of setting an intensity value
such that the values of all pixels having intensity are below the threshold value of one
phase. It is good at degree of intensity separation between the two peaks in the image

LOCAL THRESHOLDING-
LOCAL THRESHOLD is a technique which is used to convert the image which consists of
grey scale to black and white pixel images. In this technique each and every pixel images are
analysed with respect to neighbouring pixels

ADAPTIVE THRESHOLDING-
ADAPTIVE THRESOLDING is also known as dynamic thresholding. It is used to separate
desirable foreground image objects from the background based on the difference in pixel
intensities of each region.

In our paper we used adaptive threshold techniques from all given above techniques
mentioned and done its calculation using mat lab software.
Adaptive threshold technique is one of the best techniques of thresholding in field of
digital image processing and in our paper we use the adaptive threshold based on the
wavelet-shrinkage algorithm for the medical image denoising

The paper discusses about denoising of medical images such as MRI, CT, ultrasonic images
which are the input images whose denoising (noise reduction and feature and quality
preservation) and the output parameters as considered as PSNR,MSE these all operations
are done by using adaptive thresholding techniques calculated using mat lab and based on
wavelet- shrinkage algorithm

WAVELET SHRINKAGE METHOD


Wavelet is a rapidly decaying wave like oscillation with zero mean. Unlike sinusoidal
waves which extend to infinity, wavelet exists for a finite duration. They come in different
sizes and shapes. The availability of wide range of wavelets is a key strength of wavelet
analysis. We can choose the right wavelet based on our application. The novel method of
reducing image denoising is using wavelet shrinkage method.

The wavelet transform has been used to suppress noise in digital images .It has been
shown that the reduction in absolute value in wavelet coefficients is successful in signal
restoration. This process is known as wavelet shrinkage.

 As wavelets provide different scaling to localize features in our data, one can
preserve important features in original signal while removing noise. The wavelet transform
concentrates features of signal in a few large-magnitude wavelet coefficients.

In images, edges are places where the image brightness changes rapidly.
Maintaining edges while denoising an image is critically important for perceptual quality.
While traditional low pass filtering removes noise, it often smoothens edges and also
adversely affects image quality. Wavelets are able to remove the noise while preserving the
non-cognitive important features.

Wavelet thresholding is simple technique in which one coefficient is dealt with at


one time. A threshold value is set and each coefficient is compared against threshold. If the
coefficient is less than threshold, it is set to zero and if the wavelet coefficient is greater
than zero it is remained or modified. Lower values are the ones containing the noise and can
remove them. Thus lower values are removed and higher values are modified lowering the
noise. After the coefficients are thresholded , the original data of signal is reconstructed
using inverse wavelet transform.

Wavelet thresholding involves three steps.

1. A linear discrete wavelet transform


2. nonlinear thresholding Step
3. A linear inverse wavelet transforms.

You might also like