Complete Blood Count test by end-to-end Deep Learning

Project Title

Complete Blood Count test by end-to-end Deep Learning

Project Area of Specialization Artificial IntelligenceProject Summary

The CBC (Complete Blood Count) is a test which tells about the blood components’ deficiency and efficiency. It is a very important test because by this test we come to know about the diseases of blood. The problem which we are facing in past research’s is that they are done by intrusive methods which is painful and risky for the anemic patient. This project will be done by the non-intrusive method and by using deep learning which can solve that problems. The deep learning allow us to use more complex sets to extract many features in comparison with machine learning and the non-intrusive method which is free of needle that gives comfort to the patient and help to avoid the spreading of  serious diseases like HIV and AIDs that can be easily spread by the needle. One of the basic advantages of this CBC test is that the pandemic (Covid-19) which is spread across the world now a day’s can also be recognize by the WBC count which is a part of CBC test.

Project Objectives

• Development of a non-intrusive approach for complete blood count test.
• CBC estimation from non-intrusive data (available online) using deep learning models.
• Development of an Embedded system using non-intrusive sensors for data acquisition.
• Function mapping from acquired data to estimate CBC test.

Project Implementation Method

This project comprises of two phases. In phase (I) we are collecting the data online and trained that data sets on the DL (Deep learning) models. After that CBC estimation is made and then validation on test-set.
In phase (II) we will made a novel device comprises of a sensor and an embedded system. The data set which will be collected from different labs then trained on DL models and then validation of that data occurs.

 

Figure 1 Phase 1

images/Complete Blood Count test by end-to-end Deep Learning _1639952496.png

Figure 2 Phase 2

images/Complete Blood Count test by end-to-end Deep Learning _1639952497.png

Benefits of the Project

This project gives benefits to the medical department. It is advantageous to the doctors, patients, labs and to the overall society. The doctors can be benefited in such a way of easiness that just by an image of a body part they can detect the blood disease which is not time consuming. It gives comfort to the patient in such a way that the risk of blood extraction is end and the fear of needle is rectified. It allows fast accurate and painless information of blood and its constituents. In future the medication can be done by a single device which is beneficial for the society.

Technical Details of Final Deliverable

The purpose of creating this embedded system is to increase the patient’s comfort and frequency of blood testing. This non-invasive approach towards estimating CBC will provide a painless, quick and portable alternative for blood diagnosis. Sensors such as high-definition camera, red/white LEDs, phototransistors and pulse oximeter will be used to acquire data from different parts of patient’s body (fingernail bed). We will train our Deep-Learning model using Python/TensorFlow programming using Anaconda/Jupiter interface. The end delivery product will be a hardware and software integrated system working together in real-time.

Final Deliverable of the Project HW/SW integrated systemCore Industry HealthOther Industries Medical , Manufacturing Core Technology Artificial Intelligence(AI)Other Technologies OthersSustainable Development Goals Good Health and Well-Being for PeopleRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	50000
Data acquisition	Equipment	1	10000	10000
Camera	Equipment	1	5000	5000
Lighting Source	Equipment	1	2000	2000
Microcontroller	Equipment	1	5000	5000
Electronic Microscope	Equipment	1	25000	25000
Miscellaneous	Miscellaneous	1	3000	3000