Development of an AI based Android App for disease detection using blood smear

Among many other diseases that can be diagnosed using a Microscope, Malaria is one of them. According to World Malaria Report 2020, there were an estimated 229 million cases in 87 countries, and 384 000 related deaths in African region in 2019 alone. Anemia is a disease whose symptoms can be observe

Project Title

Development of an AI based Android App for disease detection using blood smear

Project Area of Specialization

Artificial Intelligence

Project Summary

Among many other diseases that can be diagnosed using a Microscope, Malaria is one of them. According to World Malaria Report 2020, there were an estimated 229 million cases in 87 countries, and 384 000 related deaths in African region in 2019 alone. Anemia is a disease whose symptoms can be observed If a patience lacks enough RBCs, and People with thalassemia may have mild or severe Anemia. Whereas increased RBCs could be one of the symptoms of Polycythemia. We have come up with an idea to develop an Android App that automates the process of Malaria Diagnosis, and counts RBCs & WBCs using Microscopic Images of Stained Blood Smear Slides. As per our research and surveys in different laboratories/hospitals, trained experts/technicians manually inspect and identify Parasites, and count blood cells on Stained Blood Smear Slides using a Light Microscope. We will be training a TENSORFLOW Model (ML/DL) using a dataset of Microscopic Images of stained Blood Smear Slides which will help automate Malaria Diagnosis and RBCs/WBCs count. These images can be acquired using a portable Digital Microscope or a Smartphone camera mounted on the eyepiece of a Light Microscope. The trained model will be converted to a TENSORFLOW lite version and embedded into Android Application for end user Interaction. With minimum amount of training of technicians to perform the diagnosis, our model will not only reduce the diagnosis process down to a matter of minutes, but it will also be cost efficient and will lower the workload on experts who otherwise manually observe the slides which may lead to false-diagnosis.

fig. 1 Concept design

Project Objectives

• Automate the Detection of Infected and Uninfected RBCs to diagnose Malaria.
• Identify and classify different species and stages of Malaria.
• Count RBCs and WBCs to observe if there is any imbalance.
• Train and test a model on TENSRFLOW (ML/DL).
• Incorporate a 100x zoom Microscope objective lens with Digital Microscope to produce results similar/close to a Light Microscope.
• Acquire images taken from Digital Microscope to smartphone for testing and training of model.
• Develop a user-friendly Android App.
• Minimize chances of False-Diagnosis.
• Collect our own labeled dataset of Blood samples from laboratories.
• Reduce the amount of time and cost required for diseases diagnosed using a Microscope.
• Provide people living in remote areas a quick and easy track for Diagnosis of diseases which requires minimum training of technicians.

Project Implementation Method

• TENSORFLOW is a platform from google used for creating Machine learning models.
• A TENSORFLOW model based on different techniques such as CNN/SVM will be trained and tested on Python using a Dataset (our collected Blood Smear slides samples along with a Dataset available online: https://bit.ly/3bVmyPc).
• The model will then be converted into a TENSORFLOW lite version which a smartphone operating system supports.
• The TENSORFLOW lite version will be embedded into Android App.
• We will develop the App using Java Programming Language in Android Studio.
• The App will acquire Images from Digital Microscope (connected using a micro-USB cable) or images captured using Smartphone camera mounted on a Light Microscope with an adaptor.
• The image acquired will be analyzed, and the results are produced and displayed on the smartphone screen.
  fig. 2 Project Implementation Flowchart
  

Benefits of the Project

Cases of Malaria are predominantly encountered in remote areas where there may be no clinics/laboratories with proper equipment (like a Light Microscope which is expensive) and experts/technicians for the diagnosis. Therefore, people have to travel to the nearest city/town with adequate clinics/laboratories for the diagnosis. Our project provides a perfect solution for that for which a technician requires minimum training for the preparation of a Slide, an inexpensive setup of Digital Microscope or a Smartphone and an adaptor for Microscope to mount smartphone on microscope’s eyepiece. Moreover, the traditional method of Malaria diagnosis and blood count manually carried out on a Microscope is not only tedious and time consuming but is also expensive (in private laboratories).

There are other methods:

• PCR (Polymerase Chain Reaction), it has higher sensitivity, but it is costly and less accurate.
• RDT (Rapid Diagnostic Test), has high sensitivity, but comes with inconsistency in results and inability to differentiate among different species of Malaria Parasites.

  Our project will provide a much faster, efficient and inexpensive way for monitoring/testing of diseases that are diagnosed using a Light Microscope. In addition to Diagnosis of Malaria our project also quantifies RBCs and WBCs whose imbalance could be the symptoms of Anemia, thalassemia and Polycythemia.

Technical Details of Final Deliverable

The final deliverable will be an Android App which will acquire images of Blood Samples from a Digital Microscope with an additional 100x objective lens or using a smartphone camera mounted on a Light Microscope with an adaptor to analyze the Blood Slides and provide results on the smartphone screen. The App will, detect and identify species and stages of malaria and provide RBCs/WBCs count.

fig. 3 Smartphone mounted on the eyepiece of a microscope.

fig. 4 Digital Microscope.

fig. 5 This is how the App might look like.

fig. 6 Screenshots of basic App interface.

fig. 7 Blood slide Image captured with a smartphone from a microscope’s eyepiece.

Final Deliverable of the Project

HW/SW integrated system

Core Industry

Health

Other Industries

IT , Medical

Core Technology

Artificial Intelligence(AI)

Other Technologies

Sustainable Development Goals

Good Health and Well-Being for People, Industry, Innovation and Infrastructure

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com