Identiﬁcation and Classiﬁcation for Diagnosis of Malaria Disease using Blood Cell Images

: Machine Learning is a subﬁeld of artiﬁcial intelligence that focuses on developing intelligent algorithms capable of learning from available data without requiring conﬆant programming, enabling them to adapt to diﬀerent environments based on current scenarios. These algorithms are crucial in making intelligent decisions and conducting thorough analyses to uncover intricate patterns concealed within the data. This ﬆudy used multiple machine-learning classiﬁcation algorithms to analyze patients' data based explicitly on input images containing parasite-infected and uninfected Malaria samples. AI techniques were utilised to measure the presence of parasites in the images. The image classiﬁcation syﬆem was designed to accurately identify malaria parasites in blood images by generating image features related to color, texture, and cell and parasite geometry. A classiﬁer based on SVM (Support Vector Machine) provided by Weka was employed to diﬀerentiate between parasite-infected and non-infected blood images. Through extensive experimentation, it was determined that SVM ﬆrategies exhibited signiﬁcant relevance, achieving a cross-validation accuracy of 99.4% in the basic diagnosis of malaria fever. This ﬁnding holds great potential in assiﬆing clinicians with accurate infection diagnoses.


INTRODUCTION
Malaria is a potentially life-threatening disease caused by parasites transmitted to humans through the bites of infected female Anopheles mosquitoes. It is prevalent in tropical and subtropical regions, particularly in sub-Saharan Africa, where it poses a significant health burden. The disease is caused by several species of Plasmodium parasites, with Plasmodium falciparum being the most deadly. When an infected mosquito bites a person, it injects the parasites into their bloodstream. The parasites then travel to the liver, where they mature and reproduce. Afterwards, they invade red blood cells, causing symptoms such as fever, chills, headache, muscle aches, and fatigue.
Malaria has become a worldwide health problem, particularly in Asian nations [1]. These environmental changes or conditions, such as temperature, rain, property use, humidity and deforestation, contribute to mosquito diseases, leading to countless deaths worldwide [2][3]. No matter the age variable countless deaths have been dropped from (estimated) 839000 in 2000 (array: 653000-1.1 million) to 438000 in 2015 (array: 236000-635000), i.e., 48% listed. In general, it's projected that malaria deaths will be diminished by 60% population increase worldwide [4]. Malaria -A Mosquito vector disorder brought on by a sting of erectile dysfunction "Anopheles" [5][6]. These nighttime mosquitoes later bite sporozites and enter into the human blood. Also, it impacts the liver first, and the developed sporozites are burst and shaped as mezosites which impacts RBC -Red Blood Cells. These mezosites are known as Parasites. A Parasite includes the nucleus & cytoplasm employed in preliminary foundation for classifying a parasite to a non-parasite [7][8]. Table 1 provides the Malaria Parasites features in table. Single-photon emission computed tomography (SPECT)" As described above, malaria-effective people are not getting proper diagnosis/healthcare facilities worldwide. There are many reasons, but here try to resolve maximum issues using images of blood cells by machine learning techniques. These problems are not sorted out at once. We need some work on it depending upon the condition of the patients' blood cell images to diagnose the patients.in this case; the doctor has to fulfil the responsibility of awareness to the people and sort out the issues according to providing results by images of blood cells. In this regard doctor more help to save the life of the person. Identification of the disease is the biggest problem in Pakistan.
Steps should be taken to prevent our urban and ruler areas from malaria by using machine learning. Automating the diagnosis process will enable accurate diagnosis of the disease and hence promises to deliver reliable health care to resource-scarce areas. Malaria disease can be categorized as uncomplicated. Malaria is a curable disease if diagnosed and treated promptly and correctly. We will use an intelligent classifier for the classification of that data.

RELATED WORK
Several strategies are recorded for the early detection of malaria disease based on the techniques of GIS 3 and the Environmental & Remote sensing techniques [11][12]. Various methods have been developed, proposed and analyzed in order to detect the malaria disease. But accuracy in detecting malaria and diagnosing it within the time is essential; otherwise, it causes the death of an individual 12-any diagnosing method needed to collect the blood samples to identify malaria disease. Two different blood films are used to identify malaria: thick and thin blood imges6. Thick blood films accept the blood samples for detection of malaria parasite density. Thin blood films accept the blood samples to identify or characterise malaria parasites. The following section gives the survey of the research that has taken place to detect malaria disease.

PROPOSED METHOD
For this research, the machine learning-based method is proposed; a complete block diagram of the proposed methodology is shown in Figure 1. Figure 1 shows that the proposed methodology is divided into two phases which are as follows:-

3.1.
Training Phase This the initial and most important phase of this reseach work in this phase the dataset acquire and converted into the feature vector for further processing. For this purpose an image base dataset of malaria parasitized and Uninfected is used . this dataset is taken from Kaggle and send it for the processing . In preprocessing layer data segmentation, moving average and in normalization form then forward in the application layer.

3.2.
Validation Phase Pre-processing, In this process, images is broke down into different segments from feature of images. Malaria Images is convert into numerous parts or fragments for rearrangements of images examination. Application layer is consist of two process prediction layer and a performance layer.
In Validation (Software model), filtered images are used for prediction; if the found disease is referred to the concerned doctor; otherwise, they are treated as a normal patient and fit considered. In the prediction layer machine learning algorithms are used to gain the data for performance. In the performance layer, find out performance and accuracy by using RMSE.

3.3.
Data Collection and Description AI depends vigorously on information that makes calculation preparation conceivable. Modern companies produce gigantic amounts of data. Despite the measure of data and information science ability we have, AI might be futile or even unsafe with poor information-gathering process set-up.

3.4.
Data Pre-processing Data Pre-processing is a system that changes crude information into a perfect informational collection. At whatever point the information is assembled from various sources, it is gathered in a crude organization, which isn't possible for the examination.

RESULTS AND OUTCOMES
The outcome illustrates minimum, maximum and average results by determining the error and accuracy of overall Malaria datasets as the results have shown us the pictures of the data. Compared to the other results mentioned here, there is always room for improvement in the system designed by humans. The thesis is based upon the pure work of the research. The system has come up with an outstanding result of 95.2% compared to the previous results of SVM. The analysis of the results has come up with one that is touching almost the same as the SVM. In Matlab, machine learning techniques apply to the same data in different sizes of data, and results of different classifiers are applied to it, and the results are the following.   Table 1.3 results here, the SVM is 99.5%, our obtained result. The best we could produce with our experiments. So much accuracy is adopted through our technique as before. It was up to 70 or 95.1%. The results are closed, but the difference in points is still the result of betterment and can be very small at times. Adopted approach, results are compared and closed at the end of time. We matched the best to produce the result of 95.2% as the linear discriminant was adopted. The earlier results were also impressive as they stand 91.5% but still the room improvement was there and it helped us to get the best and it was done.

Matlab Classifier Results
In figure 5 shows the parasite of actual positive

Figure 4: SVM Result
In figure 4 shows training results using SVM model. At the same time, the accuracy of validation results 99.4% found.
rate correct prediction of predicted class and AUC true positive class.

CONCLUSION
As there is malaria disease that needs permanent monitoring.

FUTURE WORK
In the future, system accuracy can be improved by adding some additional attributes and parameters. Also, with the increase in the number of instances, data can be analyzed using clustering techniques and association rules, and it will also improve the accuracy of the proposed model. Furthermore, using diagnostic images for disease prediction using sharp image processing techniques can also extend and improve the proposed model.

REFERENCES
[1] Angel Molina a, José Rodellar b, Laura Boldú a, Andrea Acevedo a b, Santiago Alférez c, Anna Merino a.Automatic identification of malaria and other red blood cell inclusions using convolutional neural networks. Computers in Biology and Medicine Volume 136, 104680, September 2021. [2] Md Abdur Rahim a , Ali Newaz Bahar b , Mohammad Motiur Rahman ,Automatic malaria disease detection from blood cell images using the variational quantum circuit Informatics in Medicine Volume 26, 100743, 2021.