Malaria Cell Image Classification using Deep Learning 
Jaspreet Singh Chima1, Abhishek Shah2, Karan Shah3, Rekha Ramesh4
1Jaspreet Singh Chima., Department of Computer Engineering, Shah and Anchor Kutchhi Engineering College Mumbai University, India.
2Abhishek Shah., Department of Computer Engineering, Shah and Anchor Kutchhi Engineering College Mumbai University, India.
3Karan Shah, Department of Computer Engineering, Shah and Anchor Kutchhi Engineering College Mumbai University, India.
4Dr. Rekha Ramesh, Associate Professor, Department of Computer Engineering, Shah and Anchor Kutchhi Engineering College Mumbai University, India.
Manuscript received on February 28, 2020. | Revised Manuscript received on March 22, 2020. | Manuscript published on March 30, 2020. | PP: 5553-5559 | Volume-8 Issue-6, March 2020. | Retrieval Number: F9540038620/2020©BEIESP | DOI: 10.35940/ijrte.F9540.038620
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Malaria caused by the Plasmodium parasites, is a blood disorder, which is transmitted through the bite of a woman Anopheles mosquito. With almost 240 million cases mentioned each year, the sickness puts nearly forty percentage of the global populace at danger. Macroscopic usually take a look at thick and thin blood smears to identify a disease or a cause and figure it out what weakens them. However, the accuracy depends upon smear quality and awareness in classifying and counting parasite and non-parasite cells. Manual evaluation, which is the gold standard for diagnosis requires various steps to be performed. Moreover, this process leads to overdue and misguided analysis, even when it comes to the hands of expertise. In our project, we aim at building a robust, minimized reliance of humans, sensitive model for automated analysis of Malaria. A category of deep learning models, namely Convolutional Neural Networks, guarantee especially versatile and advanced outcome with end-to-cease attribute extraction and categorization. The precision, unwavering quality, velocity and cost of the methods utilized for malaria examination are key to the diseases’ cure and ultimate eradication. In this study, we compare the overall performance of pre- trained CNN primarily based DL model as characteristic extractors closer to classifying parasite and non-parasite cells to aid in progressed sickness screening. The highest quality model layers for attribute extraction from the underlying records, is determined experimentally. The dataset has a variety of Parasite and Non-Parasite pictures of blood samples. To achieve accurate outcome, we have selected certain dominating features such as size, color, shape and cell count from the images which will help in the categorization process. Pre-trained CNNs are used as a promising tool for attribute extraction; this can be determined by the outcome of its statistical validation. Given these developments, automated microscopy could be a very good deal in the chase towards a low-priced, effortless, and dependable method for diagnosing malaria.
Keywords: Image Classification, Malaria Cell, Res Net, Deep learning, Convolution Neural Network.
Scope of the Article: Deep learning.