Annual: 2019

AP007 »
Low Cost Automated Diabetic Retinopathy Detection System
📁Digital Design
👤Parikshit Saha
 (Independent Researcher)
📅Jun 26, 2019
Regional Final

55


👀 872   💬 15

AP007 » Low Cost Automated Diabetic Retinopathy Detection System

Description

Effect of Diabetes on the eye is known as Diabetic Retinopathy which may lead to blindness in its furthermost stage. Highly skilled Doctors or Ophthalmologists use to inspect the patient's eye to detect the affectedness of this disease. But the amount of well-trained doctors is very less compare to drastically increasing amounts of patients. So proper treatment is nearly impossible. Plus also founding doctors in the remote locations is rarely possible. This Diabetic Retinopathy comes with no early symptoms. Only regular eye-checking need to be done to diagnosis it's starting affectedness. So we have planned to make an engineering-based solution making a standalone system which will do the primary diagnosis of the eye & suggest the patient the immediate precaution they need to take.

Project Proposal

1. High-level Project Description

  1. Purpose of the Design:-  Effect of Diabetes on the eye is known as Diabetic Retinopathy which may lead to blindness in its furthermost stage. Highly skilled Doctors or Ophthalmologists use to inspect the patient's eye to detect the affectedness of this disease based on some of the many detection parameters like amount of bloodvessels, microaneurysms, exudates that decides the stage of affectedness on eyes. Doctors can manually detect and diagnose the stage of diseases. But the amount of well-trained doctors is very less compare to drastically increasing amounts of patients. So proper treatment is nearly impossible. In addition, well-trained doctors' availablility in the remote locations is quite limited.
    Diabetes is also rising rapidly compared to number of well trained ophthalmologists, so if we can use an engineering based approach to detect and diagnose its effected on eye, it would be very much beneficial. Also doctors can use their precious time in research work rather than doing the boring eye checking task. This Diabetic Retinopathy comes with no early symptoms. Only regular eye-checking need to be done to diagnosis it's starting affectedness. So we have planned to make an a standalone embedded system which will do this task perfectly without help of medical or technical expert, and on the same time considering cost optimization of the final product.
  2. Application Scope:- Final outcome of this project is an cost optimized standalone embedded system which is able to acquire real-time fundus image and on the go same time detect and diagnose diabetic retinopathy’s stage and its effect on eye. It’ll also prescribe the required further steps needed to be followed as per their current eye health condition.
  3. Targeted Users:- Patients from remote locations will be the most beneficiary as they will not need to come a lot of distance just for the checking purpose of their eye health condition as the final outcome product will be of low cost and easily portable and it’ll be easily affordable by any private or government hospital.
    Furthermost the doctors will also be beneficial as they are not required to do the boring eye checking job repeatedly for a long duration. They can utilize their precious time in terms of doing innovation research work.

2. Block Diagram

3. Intel FPGA Virtues in Your Project

Here in this implementational research work we have used realtime image processing on medical image known as fundus image to detect disease stage based on features like amount of bloodvessels area, exudates area & microaneuryms area. Here the input image is of very large size. So the processing algorithm needs much more time to execute. So we need an approach for faster processing

FPGA board provides faster processing by proving us a parallel processing hardware architecture. Altera DE10 Nano SoC FPGA device’s technology feature characteristics are greatly beneficial for prototype development because of its inbuilt SoC-FPGA alongwith its HPS core & low-power ARM core processor (Dual-core ARM* Cortex*-A9 MPCore processor at 800 MHz). The Neon™ media-processing engine with double-precision floating point unit is best of realtime fast image processing operation. FPGA's reconfigurability and reprogrammability allow possible future software / hardware updates such as algorithm's improvement.

Furthermore Altera FPGA’s software platform Quartus & many a user-friendly IPs are beneficial for a novice FPGA developers in developing Altera FPGA based standard system. This DE10- Nano Development board provides a daughter board "D5M MIPI Camera Module" which alongside with a customize designed lens will be suitable for our application specific high resolution realtime fundus image capturing. Also this DE10 Nano Development platform has HDMI output connectivity features so directly connect a display unit. Which greatly meets the criteria to create a prototype of our application based system development.

I'm also working on different stand-alone SBC (Single Board Computer) Platform like CPU based Raspberry PiNvidia GPU based Jetson Nano Architectural Platform. So in future we can provide a detailed comparative detailed analysis in term of performance parameters. 

Intel® FPGA University Program Channel also provides a lot for learning different development and education boards specific tutorials & exercise alongwith their state-of-the-art software tools.

4. Design Introduction

5. Function Description

6. Performance Parameters

7. Design Architecture



15 Comments

Arindam Chakraborty
The project proposal is very strong as it's focusing on a standalone system.
🕒 Jul 05, 2019 09:02 PM
AP007🗸
Thank you. I'm mainly focusing on building of low-cost standalone compact portable system.
🕒 Jul 05, 2019 09:06 PM
JADUNANDAN NANDI
Excellent concept. Country like India where Diabetic Retinopathy is a major problem and low cost solution will benefit the most.
🕒 Jul 05, 2019 03:43 PM
AP007🗸
Thank You.
🕒 Jul 05, 2019 05:59 PM
Francisco Guerrero
Hi, Im doing a similar proposal on kaggle , the implementation over a DE 10 nano is very promising and useful, feel free to contact me for cooperation
f.guerrerol1@outlook.com
🕒 Jul 03, 2019 05:20 AM
AP007🗸
Hi Francisco Guerrero,
Kaggle Proposals used to work on Training-Test Database Image which maybe a great approach if you're having huge no of database on same type of fundus images. But using of already available Fundus Cameras are not at all an cost effective approach. So my innovation in this new designing proposal here is fully based on algorithmic approach of Computer Vision algorithms in FPGA hardware as well as designing of a low cost fundus image acquision system.
🕒 Jul 03, 2019 11:50 PM
ANIRBAN CHOWDHURY
This is a very innovative project. Diabetic Retinopathy is a very alarming problem which needs to be addressed in a proper way. The use of advanced techniques in the field of embedded system and algorithms have the potential to revolutionize the automated detection of such diseases. The proposed project has the potential to come-up with a unique solution which could benefit a large patient population.
🕒 Jul 02, 2019 05:49 PM
AP007🗸
Thank You.
🕒 Jul 04, 2019 06:29 AM
shashank kumar
nice project it will become very usefull for patient in the comming eras
🕒 Jul 01, 2019 03:28 AM
AP007🗸
Thank You.
🕒 Jul 04, 2019 06:29 AM
Doreen Liu
Do you try to use D5M? Maybe you can also refer to D8M(which is a 8 Mega Pixel Digital Camera Package ):
https://www.terasic.com.tw/cgi-bin/page/archive.pl?Language=English&CategoryNo=68&No=1011
you can get the DE10-Nano+D8M project from this link:
https://www.terasic.com.tw/cgi-bin/page/archive.pl?Language=English&CategoryNo=167&No=1046&PartNo=4
🕒 Jun 26, 2019 12:35 PM
AP007🗸
Thanks a lot.
Yes i have seen the module D8M 8MP Camera Module also. But I have written my proposal based on cost effective solution, So have mentioned here D5M Camera Module. As i guess the price will be lower & also my image processing algorithm will be sufficient to do it's task on this much resolution also.
Obiviously 8MP D8M module will be more beneficial in term of good higher quality image output.
🕒 Jun 26, 2019 02:50 PM
Doreen Liu
Fine!
🕒 Jun 26, 2019 02:58 PM
Doreen Liu
Excellent proposal!
Hope your work can bring convenience to life!
🕒 Jun 26, 2019 11:43 AM
AP007🗸
Thank you Doreen Liu.
🕒 Jun 26, 2019 12:28 PM

Please login to post a comment.