Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
This part presents the theoretical writing, related studies, theories and principle
that would serve as the theoretical basis for the researchers study and would offer
2.1.1 Fecalysis
digestive tract. Among these conditions are the infection from parasites, viruses,
be place into slides, by using applicator stick, the sample is spread in the
RBC, puss cell and parasites. Microscope is adjusted to its high power field
analyzer that can examine multiple stool samples. Its features include
7
automatic dilution, high speed mechanical mixing, automatic sample
loading and it continuously processed stool samples for about 120 samples
Belson LXL 1000 operates by putting the sample in the cup which
is well sealed and extremely sanitary and put it on the 1st slot for the sample
cups. There are 24 slots for sample cups which makes it possible to test
24 samples at the same time. There is a test card, which is used for the
examination of OBT, gut microbes, virus and etc. There are two needles
here. These needles add reagent to the sample cup and mix them
through the pipeline and put samples on the test card. And the pipeline
equipment. And they can display the result of the test on the computer
screen. There is an external camera which is used for capturing the test
8
result of the test card and the result is displayed on the screen. There is
also internal camera which is used to capture the shape, color and
Figure 2.2 Sample image of the Morphological Structure of Entamoeba histolytica Cyst
consisted of two stages, the trophozoite and cyst. The trophozoites may range in
size from 10-60 um and have irregular circular shape and a single nucleus with a
prominent nuclear border, central karysome and vacuole. Trophozoites ingest red
blood cells and absorbs their nutrients to survive. Some trophozoites transform
into cysts.
9
On the other hand, Figure 2.2 shows that Entamoeba histolytica cyst is
a number of nuclei which ranges from one nucleus in early stage and four nuclei
when mature. The nucleus itself can be describe as the ring and dot portion of the
given protozoa.
2.1.3 Algorithm
From these techniques, the algorithm that will be used in this study is selected.
10
original image and are frequently used as a preprocessing step to
Enhancement
a. Histogram Equalization
b. Median Filtering
noise such as negative exponential and salt and pepper type noise
the arithmetic mean filter in removing salt and pepper noise from an
image. The median filter can easily remove outlier noise from image
c. Contrast Correction
11
defined as stored in the structured image. The program then
computes the average of the image and removes this value from
to add back the average to the contrast corrected pixel and store
a. Edge Detection
input matrix with the Sobel or Roberts kernel. It outputs two gradient
12
operation. It can perform thresholding operation on the gradient
b. Thresholding
multi-gray level image into binary image containing two gray level
values.
c. Clustering
image.
from microscopic images. The methods used for this paper are different protozoan
parasites segmentation scheme. First, the image was first transformed into gray-
level ones and equalized color spaces using gamma equalization. Then, median-
13
mean filter and two-class edge enhancement algorithm were used to erase noises
and enhance the boundaries of protozoan parasite. K-mean algorithm was used
to cluster the pixels of each image into two major clusters which are the protozoan
parasite and the background. The largest clustered or connected pixels are
nucleuses inside it are needed to be detected. The detection scheme that they
used are composed of different methods. This includes the adaptive boundary
component detection and circular mask scoring methods. They are performed
iteratively to detect the first most significant nucleus inside the protozoan parasite.
Lastly, the height and the width of the nucleus were measured and considered as
the conditions for further detection of other nucleus. The scheme gained 96.64%
accuracy for 112 experimental protozoan parasite microscopic images but the
images. The main focus is to determine whether there exists amoeba in the image
through image processing techniques. The methods used in the study were: (1)
preprocessing stage (2) image enhancement and (3) image segmentation. In order
to prevent problems to the segmentation results, the researcher process the pre-
which includes two steps, run length enhancement and bi-group enhancement
14
was used in order to gain the complement contours. Lastly, the image
operation was applied to separate the edges of the objects and the background in
the image. The results obtained 98.73% accuracy from 75 experimental images
Network
having digital microscopic cyst images for the detection of Entamoeba species
method for the detection of several species. An algorithm was clearly presented
on how the detection from digital images could be obtained. First, a pre-processing
stage was performed. It was made possible through edge detection using multi-
flow model. A feature extraction and dimensionality reduction was also executed
through PCA and parasite image resizing in order for an easier differentiation of
one specie to another. Lastly, the classification stage was implemented through
PNN training in order to categorize and name a specie from one another. Indeed,
the algorithm presented through the Probabilistic Neural Network Toolbox from
MatLab resulted to 100% accuracy in which nine kinds of intestinal parasites were
15
2.2.4 - Red Blood Cells, White Blood Cells and Epithelial Cells Detection and
Kenneth M. aimed to develop a Matlab program that will detect and count the Red
Blood Cells, White Blood Cells and Epithelial Cells that are present on urine. This
microscopic urine test. This research is limited with detecting the RBC, WBC, and
Epithelial Cells. The researchers limited their study with the following cells
The program was developed using MatLab. Researches were done about
and white blood cells and epithelial cells to be able to made this study possible.
The RBC detection simply used circular though transform in imfindcircles function
because of its circular property. On the other hand, the process of detecting WBC
and epithelial cells both started by converting the image to grayscale using the
srgb to cmyk function. The cyan color space was extracted and image thresholding
were done. Image segmentation was implemented using edge detection and blob
detection and the binary image was generated. In filtering the detected blobs, WBC
and epithelial cells were differentiated by their shape and size. Their respective
31.8%. While WBC counting resulted to an average percent error 39%. Lastly,
Epithelial Cells Counting had an average percent error of 24.13% which is not that
large.
16
2.2.5 Microscopy Image Processing Analysis for Automatic Detection of
system and interfacing system. The first part of research methodology was the
image acquisition of Ascaris Lumbricoides Ova (ALO) and Trichuris Trichiura Ova
(TTO) parasites using a microscope camera. Next part is the image processing
extraction and classification. The best pre-processing approach has been explored
that focused on the segmentation of target object parasite cell. The captured
median filter and edge detection using canny filter. The final stage in the
obtained from the results of the image analysis and pre-processing stage. For the
Based on the result, more than 95% of accuracy was obtained for the classification
innovative of its kind where some infected blood samples are analyzed to
determine stage of the disease malaria. For pre-processing, the method that they
used is the converting of the colored blood smear image into binary image. Further
17
processes includes edge detection Harris corner detection and region selection.
Under edge detection, sobel operator was used since it is best for biomedical
emphasizes on regions of high spatial frequency that are related to the edges, it is
used to find the approximate absolute gradient magnitude at each point of an input
gray scale image. Harris corner detection was needed in order to find out corner
regions. Overall calculation says that in 97.73% cases the diagnosis of malaria is
blood with different stages of Malaria and used Matlab 7.14 software to study the
detection. Apart from detecting malaria, severity of the disease in human body can
the six zoos in United Kingdom. The purposed of the study is to distinguish the
Entamoeba dispar, and Entamoeba coli that were present in animal stool. It also
animals. The study used PCR in examining the samples. Initially, the samples were
used Minitab 15 for the evaluation of the data. It also used the QlAamp DNA stool
Mini Kit to directly extract parasite DNA from stool. The study concluded that the
18
presence of Entamoeba at multiple zoos in United Kingdom had the percentage of
28.9% based on the tested samples. The most common Entamoeba specie that
was detected is the Entamoeba coli. Through this study the detection and clear
recognized.
Immunosorbent Assay
and Dispar species. In this study the use of PCR (Polymer Chain Reaction) and
the two target species were considered to be morphologically identical from one
undergone DNA processing. On the other hand, ELISA offered serological method
in which the detection of the two species can be identify from the serum which was
the watery portion obtained from the stool samples. These two techniques resulted
to 85% accuracy from the several tests performed in the laboratory. Equipment
used were purely obtained from the laboratory and no programming tool was used.
2.3 Synthesis
The related literatures included in this study serves as references for the
researchers in developing their proposed program. The concepts and theories from the
19
The related studies give ideas for further improvement and development of the
proposed project. Through the review of related studies, the researchers found out the
differences and similarities between the other researchers work and the proposed study
software has been stated on their paper but the algorithms used were completely detailed.
The processing were basically divided into two parts. The first part focuses on the
detection of the protozoan parasite. It includes several algorithms to determine the region
of it. The second part is more on the nucleus detection inside the protozoan parasite. The
experiment shows that the proposed scheme can detect the nucleuses with indistinct
boundaries effectively.
For the study conducted by Lai et al, the researchers focus on developing an
focuses on whether there exists an amoeba in the image. Preprocessing stage, image
enhancement and image segmentation are the methods used. This study concentrates
The study performed by Beaudelaire Saha et.al, employs several techniques which
involves image edge detection, segmentation and parasite extraction with the application
of Neural Network Toolbox from MatLab. On the other hand, this study involves the same
approach of employing techniques including image edge detection and circular Hough
Transform but instead of using a toolbox from MatLab, this study focusses only upon the
implementation of several codes for the detection process. Both studies are concerned for
the protozoan cysts since its the most common stage of a protozoa that can be easily
identify and can be found from stool samples. This study focusses only upon the detection
of Entamoeba histolytica cyst and no any other species are concerned. This is because
other species are rarely found on stool samples. Lastly, both studies use MatLab as a
20
could be done. Indeed, this study offered a GUI (Graphical User Interface) for an easier
interaction and processing of results for the users which is not yet done by the previous
study.
The study of Caringal, Glenn A. De Lunas, Caroline E. and Endaya, John Kenneth
M is about the identification and counting of RBC, WBC and Epithelial cells in urine
samples. It used image processing using a computer programming software which is the
Matlab. Graphic User Interface Program is also implemented. Different codes are applied
such as codes for detection of circles, for reading and showing of the test image. There is
a certain algorithm used in order to attain the desired result: Image segmentation, Edge
Detection using Sobel Operator, Dilating Image using line Structured Element, Filling
Holes in the Image, Removing the Objects in the Border, and Eroding an Image.
WBC and Epithelial cells while the current study focuses in fecalysis and detection of
E.histolytica only. In contrast, both used and will use the same kind of method which is
image processing specifically image segmentation through the aid of MATLAB software.
Next, the study presented by K.H. Ghazali, R.S. Hadi and M. Zeehaida which is an
methodology of the study is divided into two parts decision system and the interfacing
system. The decision system was composed of image acquisition of ALO and TTO using
a microscope camera, image processing stage, image analysis and the classification
method. On the other hand, interfacing system was composed of complete automatic
camera for image acquisition, an arduino interfacing card and a high-end PC to process
and analyze the parasite images. The system capture the images of the sample using a
digital camera attached to the microscope and then directly process the detection of
parasites using the desktop computer. The samples were handled by the automatic
21
handler system through the interface with the desktop computer. The system was provided
With regards to the study of Pranati Rakshit and Kriti Bhowmik, the stages of
malaria are analyzed and can be detected. They used Matlab as their tool for processing
the sample images that they have collected. Earlier stages of malaria has been seen and
as it matures within RBC. Apart from, detecting malaria, their study can help other severe
diseases in human body and make way to diagnosis by efficient techniques of digital
image processing.
Then, for the research of Carl S. Regan, Lisa Yon, Maqsud Hossain and Hany M.
conducted through the feces as the sample and PCR as the method. The study focused
In comparison with the current study, this study tackles broad identification of
Entamoeba since it used a DNA extraction and differentiate various types of parasite.
Meanwhile the current researchers limit the scope in recognizing specifically the
software.
Lastly, the study of Patrick Gonin and Louise Trudel, detection and differentiation
of Entamoeba histolytica and Entamoeba dispar is made possible through the aid of an
in-depth specific biological method, PCR (Polymer Chain Reaction) and ELISA (Enzyme-
Linked Immunosorbent Assay). This is a comparative study that relates the efficiency of
each mentioned method. Researchers concluded that PCR is the most useful in
22
In comparison, the said previous study focusses primarily on the use of biological
method while this study uses image segmentation method through a computer
programming software which is the MatLab. This study focusses only on the detection of
23