Automatic Identification of Malaria Through Microscope Images

In accordance with the Necessity, there is an increase in demand for medical images for diagnosis application, the use of proper enhancement methods are inevitable. Medical imaging should be a challenging screening method for which effective, robust and cost-effective approaches are still to be developed. Regular screening for diseases like Malaria is necessary in order to identify the group at risk of visual impairment. Malaria parasites identification is currently done based on patient’s symptoms and parasitological testing. Conventional method have several drawbacks such as limited access to microscopy experts especially in rural area practice, restricted diagnostic facilities and costly.
This paper presents a novel automatic diagnosis of Malaria through Microscopic images by employing image processing techniques. The paper first introduces the existing systems for screening, with an emphasis on the detection methods. The proposed medical decision support system consists of automatically quantify and classify erythrocytes infected by Plasmodium falciparum (P.falciparum), vivax (P.vivax) , ovale (P.ovale) malariae (P.malariae) in thick blood smears[1][2][3] .The infected cell are classified in the thick blood smears where experimentation conducted in math work by image acquisition, image pre-processing including segmentation, classification and a combination of other image processing techniques are performed.
Automated diagnosis of malaria has received attention in recent trends. In the medical field especially through microscopic images the diagnosis of malaria is cost-effective and robust which quantify and classify the infected parasite more accurately. This paper describes a basic image processing implemented in MATLAB which detects Plasmodium falciparum parasites with microscopic images.

Malaria is a life threatening disease which leads to increase in dead rate. The diagnosis of the disease requires powerful and expensive tools unavailable for the rural area and local hospitals, where often the disease is endemic. Malaria is an infectious disease caused by the genus Plasmodium parasites, transmitted to human through the bites of infected female mosquitos. The five plasmodium species are falciparum (P.falciparum), vivax (P.vivax), malariae (P.malariae), ovale (P.ovale)[1][3][4].
The most widely used technique for determining the development stage of the malaria disease is visual microscopically evaluation of Giemsa stained blood smears. The manual analysis of slides is, however, time-consuming, laborious, and requires a trained operator. Moreover, the accuracy of the final diagnosis ultimately depends on the skill and experience of the technician and the time spent studying each slide and it has been observed that the agreement rates among the clinical experts for the diagnosis are surprisingly low[2].
Life Cycle of Malaria
The natural ecology of malaria involves malaria parasites infecting successively two types of hosts: humans and female Anopheles mosquitoes. In humans, the parasites grow and multiply first in the liver cells and then in the red cells of the blood. In the blood, successive broods of parasites grow inside the red cells and destroy them, releasing daughter parasites (“merozoites”) that continue the cycle by invading other red cells.
The blood stage parasites are those that cause the symptoms of malaria. When certain forms of blood stage parasites (“gametocytes”) are picked up by a female Anopheles mosquito during a blood meal, they start another, different cycle of growth and multiplication in the mosquito.After 10-18 days, the parasites are found (as “sporozoites”) in the mosquito’s salivary glands. When the Anopheles mosquito takes a blood meal on another human, the sporozoites are injected with the mosquito’s saliva and start another human infection when they parasitize the liver cells.Thus the mosquito carries the disease from one human to another (acting as a “vector”). Differently from the human host, the mosquito vector does not suffer from the presence of the parasites.
Causes of malaria
Malaria can occur if a mosquito infected with the Plasmodium parasite bites you. An infected mother can also pass the disease to her baby at birth. This is known as congenital malaria. Malaria is transmitted by blood, so it can also be transmitted through:

An organ transplant
A transfusion use of shared needles or syringes
Symptoms of malaria

The symptoms of malaria typically develop within 10 days to 4 weeks following the infection. In some people, symptoms may not develop for several months. Some malarial parasites can enter the body but will be dormant for long periods of time. Common symptoms of malaria include:Shaking chills that can range from moderate to severe, high fever, profuse sweating, headache, nausea, vomiting, abdominal pain, diarrhea, anemia, muscle pain, convulsions, coma, and bloody stools[10]. Laboratory test (microscopy)
Thick and thin blood smear study is the gold standard method for malaria diagnosis. The procedure follows these steps: collection of peripheral blood, staining of smear with Giemsa stain and examination of red blood cells for malaria parasites under the microscope of Thick smear. It is not fixed in methanol; this allows the red blood cells to be hemolyzed, and leukocytes and any malaria parasites present will be the only detectable elements.
However, the hemolysis may lead to distorted plasmodial morphology making plasmodium species differentiation difficult. Therefore, thick smears are mainly used to detect infection and to estimate parasitemia. Thin smear. It is fixed in methanol. Thin smears allow the examiner to identify malaria species, quantify parasitemia, and recognize parasite forms like schizonts and gametocytes[6][7][8].
Advantages:

It is an inexpensive method.
It gives the examiner the opportunity to quantify parasites and differentiate malaria species.

Disadvantages:
The diagnostic accuracy depends on quality of blood smear and equipment, abilities of the microscopist, parasite density and the time spent on reading the smear. All these may result in therapeutic delays.
Not suitable for large- scale epidemiological studies.
False positive. Defective blood film preparation may lead to artefacts that can be incorrectly regarded as malaria parasites. Sometimes, platelets also confound diagnosis.
False negative. It is associated with low parasite density or low number of fields examined by the microscopist.
Proposed Method
In this paper, an attempt has been made to implement the malarial diagnosis algorithm that has been implemented, tested and evaluated on a MATLAB platform, by mobile microscope.
An automated system that is capable of providing minimum human intervention and a reliable result needs to be designed. In general, mobile microscope offers an ideal platform for creating a field-based, modular polarized microscope.
In pre-processing, images were first converted to HSV image and then filtered to remove noises. For segmentation stage, a threshold for each image was calculated using Otsu method. Further, dilation and erosion were performed to completely remove background elements. In the classification stage, images were classified based on the number of infected red blood cell detected. Testing performed using images acquired from conventional microscope [8][9][10].
Methodology
Proposed system block diagram
The proposed system is designed to provide solution for the drawback of existing system. The block diagram for proposed system is shown in Figure 3.1. The description of each block is explained below.
Microscopic Image Processing
Microscopic images are captured from digital camera that is fitted with microscope. A number of images are collected. The microscope image acquisition is now processed by image processing. The image processing is carried out by fallowing steps.
Steps involved in image Processing
Image acquisition
Dried thin blood films infected with plasmodium of various stages were first fixed with methanol. Then, the thin blood films were dipped with alkaline buffer at pH7.2 and stained with Giemsa at ratio of 10%. Images were viewed under mobile microscope and acquired in mobile camera. Images were stored in JPEG format at resolution of 764 x 574 pixels.
Pre-processing
For this stage, four different processes are applied to the images. It clearly shows that red blood cells infected with Plasmodium of various stages can be differentiated from its background as it possesses a higher intensity level.
The original RGB image was converted to HSV image to extract the intensity value. Next, for filtering techniques, a contrast enhancement filter created from the negative of Laplacian filter is chosen to sharpen the image. It works by subtracting the blurred images from the original images. Median filter is then applied to reduce “salt and pepper noise” and to help preserve edges. Median filter is a nonlinear filtering operation. Wiener filter with size 5×5 is also applied to remove the Gaussian noise. As a result, each Red Blood Cell (RSC) area and boundary becomes clearer. Histogram stretching is then performed in order to adjust the contrast or intensity values of the image.
Segmentation
In segmentation process, It is one of the most difficult tasks in image processing as it determines the failure or success of the next classification process. Segmentation process followed in this project Otsu Method thresholding Negative Image transformation: Image dilation, Image erosion, Clear Border, Hole Filling of segmentation process In this paper, Otsu method is used to find a local threshold so that the pre-processed image can be converted to binary image. The main principle of this OTSU method is to cluster the object and background element as close as possible and to reduce their intersection.
It is consider fast as it operates directly on grey level histogram. According to Otsu method is capable of finding a local threshold value that can minimize the weighted within-class variance. The complement of the image is then calculated and transformed by using the negative image function. The image is Binary image using OTSU method. Canny edge detection method is also employed.
Feature extraction
It is a type of dimensionality reduction that efficiently represents interesting parts of an image as a compact feature vector.
Classification

Image classification analysis the numerical properties of various image features and organizes data into categories. Finally the infected cells are identified.
Images are collected from offline dataset. The collected image is given as input for image pre-processing.

After image Acquisition it is pre-processed by pre-processing technique where L*a*b colour based segmentation, image sharpening, image filtering and image stretching are performed.
The segmented image is obtained by otsue method thresholding and color based segmentation.
TABLE I
Extraction of feature from segmented image.type of smear Thick smear malaria parasites P.Falciparum Stage Gametocyte

Contrast 0.4731
Correlation 0.6945
Energy 0.8271
Homogeneity 0.9725
Mean 1.0507
standard_deviation11.2335
Entropy 0.1273
Rms1.1366
Variance 124.2963
Smoothness 0.9999
Kurtosis 127.6960
Skewness11.0048

The affected cell in blood smear is identified by using image processing which is experimented in matlab.
Conclusion
The image processing techniques is able to automatically detect the presence of affected cell in red blood cell. It was accomplished using the chosen Otsu’s thresholding technique. Further, the extracted features are taken to train classifiers where diagnosis is made by determining whether the image is infected or not.
The proposed medical decision support system consists of automatically quantify and classify erythrocytes infected by Plasmodium at various stages in thick blood smears where experimentation conducted in math work by image acquisition, image pre-processing including segmentation and classification techniques.
References

Apurba paul, Ponnan padmapriya , Vasant natarajan, 5 May 2017, “Diagnosis of malarial infection using change in properties of optically trapped red blood cells” Department of Physics, Indian Institute of Science, Bangalore, India ,Artical,.
M Bister 2012, “Increasing the speed of medical image processing in MATLAB”, Biomedical Imaging and Intervention Journal, pp.2-12. Casey W. Pirnstill, and Gerard L. Coté, “Malaria Diagnosis Using a Mobile Phone Polarized Microscope”, This article has been cited by other articles in PMC. Sci Rep. 2015;5:13368.Published online 2015 Aug 25. doi:  10.1038/srep13368 PMCID: PMC4548194.
Charles B.Delahun,t Courosh Mehanian,Liming Hu,Shawn K. McGuire,Cary R. Champlin ,Matthew P.Horning,Benjamin K. Wilson, Clay M Thompon 2015 “Automated Microscopy and
Machine Learning for Expert-Level Malaria Field Diagnosis” , Global Humanitarian Technology Conference ,IEEE .
Corentin Dallet †, Saumya Kareem, Izzet Kale 2014 “Real Time Blood Image Processing Application for Malaria Diagnosis Using Mobile Phones”,IEEE Conference 978-1-4799-3432-4/14/$31.00.
HardikPandit ,Dr. D M Shah 2011“Application of Digital Image Processing and Analysis in Healthcare Based on Medical Palmistry”International Conference on Intelligent Systems and Data
Processing ICISD 2011 Special Issue published by International Journal of Computer Applications (IJCA),pp.56-59.
S.Kareem,I.Kale,R.C.S Morling 2012 “ Automated P.falciparum Detection System for Post-treatment Malaria Diagnosis Using Modified Annular Ring Ratio Method” ,14th International Conference on Modelling and Simulation.
Lu´ıs Rosadoa,∗, Jos´e M. Correia da Costab, Dirk Eliasa, Jaime S. Cardosoc 2016 “Automated detection of malaria parasites on thick blood smears via mobile devices”,International Conference On Medical Imaging Understanding and Analysis 2016, MIUA 2016, 6-8 July .
Kaushiki Roy, Ratnadeep Dey, Debotosh Bhattachmjee, Mita Nasipuri , Pramit 2016“GhoshA Smart Phone Based App For Automated Segmentation and Counting of Platelets” ,3rd Int’l Conf. on Recent Advances in Information Technology I RAIT.
Satoshi Kawata and YoshikiIchioka 1990 “Iterative image restoration for linearly degraded images. II. Reblurring procedure,” Journal of the Optical society of America(JOSA),vol.70,pp.768–772.
Wahyu Andi Saputra1, Hanung Adi Nugroho2, Adhistya Erna Permanasari3 2016 “Toward Development of Automated Plasmodium Detection for Malaria Diagnosis in Thin Blood Smear Image: an Overview”International Conference on Information Technology Systems and Innovation (ICITSI) Bandung – Bali.