Use of Images as an Input

Published: 2021-09-15 09:15:09
essay essay

Category: Learning, Biology

Type of paper: Essay

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Hey! We can write a custom essay for you.

All possible types of assignments. Written by academics

How the images are being used as an input with their characteristics for the implementation of the entire approach The technical details of the procedures used in this research are listed. The experiment in the research work is being performed on the grayscale ultrasound images of a human gallbladder. It details the methodology platform and the functions being used with every step of evaluation in the proposed algorithm is discussed in detail in this chapter.
Experimental Data
As an experimental data, the methodology uses real time gray scale images that have been collected from various ultrasound centers and are evaluated through the proposed step of processing for better visualization and noise reduction. The goal of the methodology is to produce an image with significant information and minimum entropy on the basis of feature selection.File formats and image size
The images being used for the approach are in JPG format, and have been taken with variable sizes and variable dimensions, in order to give the methodology an interactive and non-sophisticated approach. A random set of the images, with their detailed properties including image size, and dimensions has been taken.
The sample images taken are found to be visually appearing similar to each other but with the varying characteristics of their image sizes, contrast levels and entropy values. In order to avoid an unbalancing condition, the images with critical abdomen diseases, and the images that have been taken under the poor lighting conditions are excluded.
Source of data
The images are personally collected from the local ultrasound centers in Udhampur. The application of proposed methodology is related to the analysis of ultrasound images of the gallstone patients in order to analyze their gallbladder medical imagery reports. The study was carried out on diagnosed, pre- operative and symptomatic patients of Cholelithiasis (Gallstone disease). For the purpose of medical imagery, the patients were supposed to be in supine position. However, those having gallbladder and pancreatic carcinoma, emergency operations, patients having age
Ultrasound images
According to the ultrasound is one of the most widely used medical diagnostic tool used for imaging. Dr. Karl Theo was the first Austrian neurologist to apply ultrasound for the purpose of medical imaging. The ultrasound images are homographic that can be acquired in real time formed from a pulse echo approach. The sound waves of frequency greater than 20 KHz (exceeding the upper limit of the human audible hearing), that is, between 1-20 MHz, being used for medical purpose from the Piezoelectric crystals [30] are allowed to penetrate the body tissues.
Among these waves, some of the waves are reflected back to the transducer as echo signals while some of them are continued for deeper penetration inside the body. The echo pulses that are returned back, thus forms the image. The amount of the echo pulses that are returned after hitting a body tissue depends on the type of tissue that it hits. This tissue property is called as acoustic impedance. The air containing organs (like lung) have lowest acoustic impedance while the denser organs (like bones) have the higher values of acoustic impedance.
Tool used
For the processing and analysis of the data images, the MATRIX LABORATORY (MATLAB) version (R 2015) 32 bit MATLAB, a high level language that serves to provide an interactive environment is used. It has enabled to perform the intensive computational tasks in more easier and faster way with high accuracy.
The Method
The method and the plan of work being used for the research work is presented through a flow chart in chapter 3, with total ten steps. These steps are being discussed in the following section.
Step1: Load Data
The user browses an image in .jpg format or .tiff format from the image directory. All the test images are taken in gray scale. However, for any rgb image, this step also allows converting rgb image into a grey scale image and thus the image is referred as a Sample Image or test image.
Step2: Removing Label and Extracting Region of Interest (ROI)
For the purpose of removing labels present in the ultrasound images, a mask is prepared and is called as pre-saved mask. This mask is superimposed on the sample image for the purpose of convolution. As the convolution between the mask and the sample image is performed, the non mask area will receive the zero weight after convolution and thus, it is discarded. This step is called as Removal of the label.
However, for the purpose of ROI extraction, a Gaussian filter is created with hsize [5 5] and sigma = 1.7. The Gaussian filter is a well known smoothing filter for spatial and frequency domains [11, entropy res paper, S. A Jameel et al., 2015]. According to the authors, the Gaussian operation with zero mean value and σ as the standard deviation in two- dimension is given as-G (p, q) = 1/(2Пσ
2 ) e
2)) Eq. (4.4.1)
The authors in [1, ref. of entropy, Marr et al., 2002] have declared the Gaussian filter as the best filter for image smoothing which has given the encouragement of using Gaussian filter in the proposed methodology. The authors have accomplished a second order derivative of Gaussian which is being used by applying laplacian of Gaussian function as a Gaussian filter. It is given by-∆2 G (p, q) = d
2 )G (p, q) + d
2 G(p, q) = (p
2+ q
2- 2σ
6 ) e
2+ q
2))) Eq. (4.4.2)
Towards, the next step of ROI extraction, a classical un- sharp masking filter is used. According to the [12, Ref of entropy, M. Obulesu et al., 2012], the un- sharp masking filter enhances the high frequency edges of the image. This process subtracts a un- sharp version of the image from an original image. For this purpose, if x is an input image and y is the result of linear low pass filter then, ν = y+ ϒ (x – y) Eq. (4.4.3) Where ϒ is the gain and is a real scaling factor. In this way, the resulting images are the images with region of interest and sharpened features. Therefore, they are called as EROIS images (Extracted Region of Interest Sharpened Images).
Step 3: To Adjust the Global Contrast of an Image
This step allows manually adjusting the contrast of a sample image. The proposed approach has given a choice of five lower contrast steps and five higher contrast steps with (0.05) as a change per step value to the user for adjusting the contrast according to the application. However, for the purpose of methodology, a contrast level of 0.30 and a brightness level of 0.50 are chosen. And after adjusting the global contrast of an image, the image will be called as the Ready Image. The original histogram along with the transformed histogram which is attained after adjusting the global contrast are also plotted in this step with input intensity on the x-axis and output intensity on the y-axis.
Step 4: To check the Global Contrasted Image
With this step, the CLAHE, Contrast Limited Adaptive Histogram Equalization is performed, this process divides the image into small regions, called as tiles and thus performs the contrast enhancement of each tile by stretching the histogram limit along with readjusting the values such that the image processing in the histogram equalization limit is performed.
Step 5: To apply a Median Filter
Standard median filter is used in this step as if there is N no. of odd elements in a group, the middle element is resumed as a median, used within any window size, S. However, if S is also an odd number, this window tends to move along with set of all sampled values for every position. The median is calculated and is re- written at the same position of the output pixel. The proposed methodology has used standard median filter (SMF), in order to remove the speckle noise present in the images. The median filter is found to be performing best to preserve the edges of filtered signals [15, 16 in ref of entropy, 2009, 2017]. In order to remove the noise, the median filter is used with the following parameters: n iterations = 2 kappa = 10 lambda = .25 m=3 and clip = 0.5.

Warning! This essay is not original. Get 100% unique essay within 45 seconds!


We can write your paper just for 11.99$

i want to copy...

This essay has been submitted by a student and contain not unique content

People also read