Scribd
Upload
46 views24 pages

Malaria Parasite Detection in Thin Blood Smear Images Using Deep Learning

This document presents a method for using deep learning to detect malaria parasites in thin blood smear images. It discusses using convolutional neural networks (CNNs) and residual networks to classify images into parasitized or uninfected classes. The methodology implements CNNs and ResNet on a dataset of over 27,000 blood smear images. Classification reports show the CNN achieving 89% accuracy and ResNet achieving 97% accuracy. Hardware and software requirements for running the deep learning models are also outlined. The study demonstrates that deep learning can effectively classify malaria in blood smear images.

Uploaded by

Rouf Ahmad
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
46 views24 pages

Malaria Parasite Detection in Thin Blood Smear Images Using Deep Learning

This document presents a method for using deep learning to detect malaria parasites in thin blood smear images. It discusses using convolutional neural networks (CNNs) and residual networks to classify images into parasitized or uninfected classes. The methodology implements CNNs and ResNet on a dataset of over 27,000 blood smear images. Classification reports show the CNN achieving 89% accuracy and ResNet achieving 97% accuracy. Hardware and software requirements for running the deep learning models are also outlined. The study demonstrates that deep learning can effectively classify malaria in blood smear images.

Uploaded by

Rouf Ahmad
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 24

MALARIA PARASITE DETECTION IN

THIN BLOOD SMEAR IMAGES


USING DEEP LEARNING

PRESENTED BY:
ROUF AHMAD TANTRAY
1702CUKmr24
(DEPARTMENT OF IT)
Contents

• Introduction
• Bottleneck in Malaria Diagnosis
• The Problem statement
• Why Deep Learning
• Dataset
• Image Analysis
• Methodologies
• Implementations
• Classification Reports
• Hardware/Software Requirements
• Conclusions
• References
Introduction

• Malaria is a disease and is caused by the bites


of infective female anopheles mosquito.
• It is a vector-borne disease.
• It is caused by the protozoan parasite of the
genus plasmodium.
• Generally, affects the human liver and Red
Blood Cells.
• Nearly 4 lakh deaths per year.
Fig.1 shows indigenous malaria cases in countries
and their status by 2017.(Source WHO ,malaria
report 2018)
A Bottleneck in Malaria Diagnosis

• Microscopic examination of blood is the best


known method for diagnosis of malaria. A
patient’s blood is smeared on a glass slide and
then stained with a contrasting agent that
facilitates identification of parasites within
RBCs. A trained clinician examines 20
microscopic fields of view at 100 X
magnification, counting RBCs that contain the
parasite out of 5,000 cells(WHO protocol).
The Problem Statement
• Manually counting 5000 cells is a slow
process. This can easily burden clinic staff,
especially where outbreaks occur. Therefore, it
is needed to determine how image analysis
and machine learning could reduce the
burden on clinicians and help prioritize
patients.
Why Deep Learning
• It skips the feature computation step and
sometimes even the segmentation step.
• Deep learning does not require the design of
handcrafted features.
• New algorithms are able to diagnose disease
more accurately than expert physicians.
• To reduce the burden for microscopists in
resource-constrained regions and improve
diagnostic accuracy.
Dataset
The dataset consists of 27,588 images belonging
to two separate classes:
I. Parasitized: This implies that malaria is in the
cell.
II. Uninfected: Meaning there is no evidence of
malaria in the cell.
The dataset is provided by National Library of
Medicine(NLM).The link is given
https://ceb.nlm.nih.gov/repositories/malaria-
datasets/
Image Analysis
Methodology
Implementation with CNN
Classification Report Using CNN
Class Precision Recall F1-score Support

0 0.92 0.80 0.86 30

1 0.86 0.95 0.90 40

Micro average 0.89 0.89 0.89 70

Macro average 0.89 0.88 0.88 70

Weighted 0.89 0.89 0.88 70


average
ResNet
• They are residual networks.
• It eases the training of networks which are
substantially deeper.
• Deeper CNN models are difficult to train.
• It also performs better than CNN over larger
number of features.
Project Structure
Methodology
Implementation
Graphical Representation
Classification Report of ResNet

Class Precision Recall F1-score Support

Parasitized 0.98 0.96 0.97 2661

Uninfected 0.96 0.98 0.97 2620

Micro average 0.97 0.97 0.97 5281

Macro average 0.97 0.97 0.97 5281

Weighted avg 0.97 0.97 0.97 5281


Software Requirements
• Keras: Keras is a deep learning framework.
Installing Keras with the TensorFlow backend.
• NumPy & Scikit-learn: If you know the Keras
installation procedure, these packages for
numerical processing and machine learning will be
installed along with it.
• Matplotlib: The most popular plotting tool for
Python. Once you have your Keras environment
ready and active, you can install via pip install
matplotlib.
• imutils: package of image processing and deep
learning convenience functions can be installed via
Hardware Requirements
• Requirement of higher processors like i5 and
i7.
• 8GB RAM needs to be installed.
Conclusions
• This study has shown that malaria parasites can be effectively
recognized on thin blood smears using a Convolutional Neural
Network (CNN) and Residual Networks.
• The end goal of this project is to develop a framework that
can help researchers to automatically classify cells into 2
classes i.e. Either belongs to infected or un-infected from a
field of view image and identify features differentiating the
infected stages.
References

• WHO. Malaria microscopy quality assurance manual-version 2. World Health


Organization; 2016. World malaria report 2016. World Health Organization; 2016.
3. Tek FB, Dempste.
• Z. Liang, A. Powell, I. Ersoy, M. Poostchi, K. Silamut,K. Palaniappan, P. Guo, M. A.
Hossain, A. Sameer, R. J.Maude, J. X. Huang, S. Jaeger, and G. Thoma, Cnn-
basedimage analysis for malaria diagnosis, in: 2016 IEEE InternationalConference
on Bioinformatics and Biomedicine 1(BIBM), (Dec 2016), pp. 493–496
• Mahdieh Poostchi, Kamolrat Silamut, Richard Maude, Stefan Jaeger, George
Thoma, Image analysis and machine learning for detecting malaria, Translational
Research (2018), https://doi.org/10.1016/j.trsl.2017.12.004.
• G. Gopakumar, V. K. Jagannadh, S. S. Gorthi, andG. R. K. S. Subrahmanyam, Journal
of Microscopy 261(3),307–319 (2016)
• Bibin D, Nair MS, Punitha P. Malaria parasite detection from peripheral blood
smear images using deep belief networks. Int J Appl Eng Res 2017;5:9099–108.
THANK YOU

You might also like