29-09-2016, 04:12 PM
1456898428-bloodreport.docx (Size: 367.17 KB / Downloads: 6)
ABSTRACT
Blood cancer disease is one of the leading causes disease of death among men in
developed and developing countries. Its cure rate and pre-diagnosis depends mainly
on the early detection and diagnosis of the disease. In order to conserve the life of
the people who are endured by the Blood cancer disease, it should be pre-diagnosed.
So there is a demand of pre-diagnosis method for Blood cancer disease which should
provide superior results in the form of percentage.
In this method we illustrate a process to classify the microarray gene expression data
based on their types blood sample images using data mining and image processing techniques.
The proposed Blood cancer prediagnosis system is a combination of FFBP
Neural Network grouping with Statistical Approach and Fuzzy Inference System.
The ultimate objective of the system is to solve the drawbacks in dimensionality
reduction as they have a direct impact on the robustness of the generated fuzzy rules
system. Consequently, the goal is to generate fuzzy rules based on dimensionality reduced
data. Then the risk factors of the dimensional concentrated dataset are given to
the Feed Forward back Propagation Neural Network to accomplish the training process.
In the testing practice.
Keywords — Data Mining, Feed Forward Back Propagation Neural Network, DNA
Microarrays, Cancer Diagnosis
INTRODUCTION
It is a connective tissue consisting of cells suspended in plasma.From the iden- tication
of blood disorders, it can lead to classication of certain diseases related to blood
cancer. In this system we describes a preliminary study of developing a detection of
leukemia(cancer) types using microscopic blood sample images. Analyzing through
images is very important as from blood sample images, diseases can be detected and
diagnosed at earlier stage. From there, further actions like detection,controlling, monitoring
and prevention of diseases can be done.
Blood’s major functions are to transport various agents such as oxygen(O2), carbon
dioxide(CO2), nutrients, wastes, and hormones. Blood cells are composed of erythrocytes
(RBC), leukocytes (WBC) and thrombocytes (platelets). The most abundant small
reddish cells are erythrocytes and called red blood cell. The conventional device used
to count blood cells is the hemocytometer. It consists of a thick glass microscope slide
with a rectangular indentation creating a chamber of certain dimensions. This chamber
isattched with a grid of perpendicular lines.
It is possible to count the chamber of cells in a specific volume of fluid, and calculate
the concentration of cells in the uid. To count blood cell, physician must view
hemocytometer through a microscope and count blood cells using hemocytometer.
The organized in order and are of equal size. Cancer cells are dierent than normal
cells. They are in dispersed order, their sizes are dierent and they are not structured
well. The need is to automate this process to make the cancer diagnosis ecient and fast
with the use of state of the art technology [3].
1.1 Aim
• In this project we are going to implement a new and efficient way to detect the blood
1
Blood cancer detection
cancer with the help of microscopic images. Detection based services can prove important
in detection of blood cancer.
• System help for aware the patient and for generating the blood cancer report.
• We are using latest technique like feed forword neural network, fuzzy neural network
for the successfull detection of blood cancer.
1.2 Objective
• The main objective of this system give the result in percentage using image processing
technique.
• Finding more accurate result by using large data.
• It should pre-diagnosis so there is a demand of pre-diagnosis method for blood
cancer disease which should provide superior results.
• This system gives efficient and fast result.
Chapter 2
LITERATURE SURVEY
There are many on going researches in the eld of medical diagnosis. Bayesian networks
(BN) plays an important role in medical diagnostics. It is used to represents
conditional dependencies among the random variables. It computes the probability of
the occurrence of the various diseases when the symptoms are given. The KNN is the
simplestclassiers and is used in predicting diseases. ANN also called NN consists of
a neurons that they are interconnected. They are represent relationships between inputs
and outputs. Patient is assigned to one of the classes of diseases with this network. Articial
NN is good in identifying diseases and does not need any details of how to recognize
as it learns by example. It is easy to maintain and has good capacity and good computational
power with good accuracy. Back propagation is used to train articial NN. It is a
supervised learning method [6].
Our simple technique utilizes a minimally invasive sample collection procedure and
an automated improve detection of cancer tumors. We provide the physician an additional
novel tool for cancer detection that may be used as part of the common diagnostic
procedure and reduce the use of expensive, complicated, and painful examinations. This
novel low-cost method may reduce the number of uncomfortable tests such as mammography,
colonoscopy, computed tomography, and MRI examinations which a patient
must endure and limits his/her exposure to radiation. Such improvements in detection
will allow for better and more ecient treatment of cancer and are expected to reduce
mortality [7].
This research involves detecting the types of leukemia using microscopic blood sample
images. The system will be made by using features in microscopic images by examining
changes on texture, geometry, colors and statistical analysis as a clas- sier input.
The system should be ecient, reliable, less processing time, smaller error, high accuracy,
cheaper costs and must be robusts towards varieties that exist in individual, sample
collection protocols, time and etc. Information extracted from microscopic images of
3
Blood cancer detection
blood samples can benet to people by predicting, solving and treating blood diseases
immediately for a particular patient [5].
The various Neural Networks that have been used in successful classication of medical
data for various disorders. Examples include - Feed Forward Neural Network, Radial
Basis Function (RBF) Network, Kohonen self-organizing network, Fuzzy Neural Network,
Probabilistic Neural Network.For the successful detection of Blood Cancers [3].
In previous, there has been an exponential aim in the classification of medical diseases
images by different classifiers and algorithm.Various algorithms have been used to
classify the medical images. Effective medical image can play an vital role in aiding in
diagnosis for healthcare to students by explaining with this image will help them in their
studies as well. Data mining is the process of finding meaningful new co-relations. Pattern
and trends by shifting through large amount of data can be stored in the data base.
From the beginning the algorithm are used for testing the existence of a natural grouping
tendencys in data acquiration and most of them being based on arguments coming
from mathematical statistic and heuristic graphical techniques. These systems enhance
the classification process to be more accurate. Many techniques such as algorithms and
classifies are used for the purpose of medical image classification fig. shows various
approaches in image processing [4].
Medical imaging has become one of the most important visualization and give the
meaning in biology and medicine over the past decade. This time has witnessed a
tremendous development of new, strong instruments for detecting, storing, transmitting,
analyzing, and displaying medical images. This has led to a large growth in the application
of digital image processing techniques for solving medical problems. The most
challenging aspect of medical imaging contribute in the development of integrated systems
for the use of the clinical sector. Design, implementation,verification and validation
of complex medical systems require a tight interdisciplinary collaboration between
physicians and engineers. Main objective of analyzing through images is to collect information,
detection of diseases, diagnosis diseases, monitoring and evaluation.At the
moment, identification of blood disorders is through visual inspection of microscopic
blood cells images. From the identification of blood disorders, it can lead to classification
of certain diseases related to blood cancer. One of the most feared by the human
disease is cancer. Leukaemia is a type of blood cancer, and if its detected late, it will
result in death. Leukemia occurs when a lot of abnormal WBC’s produced by bone
marrow. When abnormal white blood cells are a lot, the balance of the blood system
will be disrupted. The existence of abnormal blood can be detect when the blood sample
is taken and illustrated by haematologists. Microscopic images will be inspected
visually by haematologists and the process is time consuming and tiring. The process
wants human expert and prone to errors due to emotion disturbance and human physical
N4
Blood cancer detection
capability having its own limit. Moreover, it is difficult to get regular results from visual
monitoring [4].
Disadvantages Of Existing System:
1. It is very slow.
2. Gives output only Yes or No.
3. Harder to use and implement.
N5
Chapter 3
PROBLEM STATEMENT
In previous days there are not a such system that will find the blood cancer in the percentage,
the previous project they are only going to say that the patient is infected or not but in this system
we are going to display the percentage of cancer and how to diagnosis it in the better way. In
this system we are using the latest technique that is Feed Forward Back Neural Network and the
Fuzzy Interference System so,the accuracy of result is more and trying to classify the blood In
this system. genes in the human blood having the large information which can be used to study
any disease in depth, DNA micro array are used to study the information obtained from genes.
6
Chapter 4
PROJECT REQUIREMENTS
4.1 Hardware Requirement
1. Processor - 1.5GHZ Intel Pentium.
2. RAM - 1GB of system memory(2GB recommended).
3. Hard Disk - 1GB.
4.2 Software Requirement
1. Operating System - windows XP professional, Window 7.
2. Language - php.
3. Backend - Mysql.
4. Editor - Text.
5. Server - Xampp -win 1.8.0.vc9.
6. Output - Chrome,Mozila Firefox.
7
Chapter 5
SYSTEM ANALYSIS PROPOSED
ARCHITECTURE
5.1 Proposed System Architecture
5.1.1 Block Diagram
Proposed System Architecture for Blood Cancer Detection is as shown in below
figure.In Early the dimensionality of given datasets can be reduced by using Statistical
Analysis and fuzzy inference system. After performing this reduced Dimensionality
dataset can be gives as input to the pre-diagnosis stage