BEye: Software Implementation and Hardware Acceleration of Retinal Vessel Segmentation for Diabetic Retinopathy Screening Tests
0 likes294 views
The document discusses an automated retinal vessel segmentation algorithm designed for the screening of diabetic retinopathy, highlighting its 92.93% accuracy and a performance rate 172 times faster than traditional methods. It emphasizes the prevalence of diabetic retinopathy, which affects a significant portion of individuals with diabetes, and proposes future enhancements including the use of convolutional neural networks for improved detection. Financial implications are also noted, with diabetic retinopathy contributing to global blindness and substantial economic costs.
![5
[1] https://nei.nih.gov/health/diabetic/retinopathy
Diabetes and Diabetic Retinopathy
[1]
Approximately 1 in 3 people living with diabetes have some degree of
diabetic retinopathy and 1 in 10 will develop a vision impairment
form of the disease [2]
[2] Yau, Joanne WY, et al. "Global prevalence and major risk factors of diabetic retinopathy." Diabetes care (2012): DC_111909.
53
Type One Diabetes Type Two Diabetes
Diabetic Retinopathy
90%
Diabetic Retinopathy
60%](https://image.slidesharecdn.com/irenebeyestanforddef-180529212817/85/BEye-Software-Implementation-and-Hardware-Acceleration-of-Retinal-Vessel-Segmentation-for-Diabetic-Retinopathy-Screening-Tests-5-320.jpg)
![7
7Screening Test
But sometimes waiting list are too long, and it is possible to
wait from 4 months up to more than a year[3]
[3] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3553639/pdf/bjgp63-e134.pdf
12-24 months
For diabetic patients
6-12 months
For diabetic retinopathy patients
3-6 months
For diabetic retinopathy patients
in the proliferative stage
Frequency
[months]
Guidelines for diabetic retinopathy screening
5](https://image.slidesharecdn.com/irenebeyestanforddef-180529212817/85/BEye-Software-Implementation-and-Hardware-Acceleration-of-Retinal-Vessel-Segmentation-for-Diabetic-Retinopathy-Screening-Tests-7-320.jpg)
![10Results 108
[4] https://www.isi.uu.nl/Research/Databases/DRIVE/
Accuracy
Pixel comparison between manual
segmentation and Beye
= 92.93% [4]
TP + TN
Tot
Accuracy =](https://image.slidesharecdn.com/irenebeyestanforddef-180529212817/85/BEye-Software-Implementation-and-Hardware-Acceleration-of-Retinal-Vessel-Segmentation-for-Diabetic-Retinopathy-Screening-Tests-10-320.jpg)
![11Results 11
[4] https://www.isi.uu.nl/Research/Databases/DRIVE/
Accuracy
Execution Time
At the same time, our implementation performs 172 times faster
respect the state of the art (0.068 s)
Pixel comparison between manual
segmentation and Beye
= 92.93% [4]
TP + TN
Tot
Accuracy =
8](https://image.slidesharecdn.com/irenebeyestanforddef-180529212817/85/BEye-Software-Implementation-and-Hardware-Acceleration-of-Retinal-Vessel-Segmentation-for-Diabetic-Retinopathy-Screening-Tests-11-320.jpg)
![12Results 12
[4] https://www.isi.uu.nl/Research/Databases/DRIVE/
Accuracy
Execution Time
At the same time, our implementation performs 172 times faster
respect the state of the art (0.068 s)
Pixel comparison between manual
segmentation and Beye
= 92.93% [4]
TP + TN
Tot
Accuracy =
VS.
Intel Core i7-6700 CPU Xilinx Zynq-7000
At the same time, our hardware implementation performs
6 times faster respect our software implementation
8](https://image.slidesharecdn.com/irenebeyestanforddef-180529212817/85/BEye-Software-Implementation-and-Hardware-Acceleration-of-Retinal-Vessel-Segmentation-for-Diabetic-Retinopathy-Screening-Tests-12-320.jpg)
![11
[12] Resnikoff S, Pascolini D, Etya’ale D, et al. Global data on visual impairment in the year 2002. Bulletin of the World Health Organization 2004; 82: 844–51.
[13] The Global Economic Cost of Visual Impairment: Summary Report April 2010
It is estimated that Diabetic Retinopathy accounts for 4.8% of the number of cases of
blindness worldwide [12]
The financial cost of visual impairment worldwide was US$2,954 billion in 2010 [13]
Conclusion 1310](https://image.slidesharecdn.com/irenebeyestanforddef-180529212817/85/BEye-Software-Implementation-and-Hardware-Acceleration-of-Retinal-Vessel-Segmentation-for-Diabetic-Retinopathy-Screening-Tests-15-320.jpg)
![10Conclusion
Thank you for your attention
https://www.facebook.com/beye.project/
https://www.slideshare.net/BEye_project
https://twitter.com/BEye_NECSTLab
irene.fidone@mail.polimi.it
It is estimated that Diabetic Retinopathy accounts for 4.8% of the number of cases of
blindness worldwide.[12]
The financial cost of visual impairment worldwide was US$2,954 billion in 2010[13]
13
May 29th, 2018
Da Vinci Conference Room @Stanford OTL
10](https://image.slidesharecdn.com/irenebeyestanforddef-180529212817/85/BEye-Software-Implementation-and-Hardware-Acceleration-of-Retinal-Vessel-Segmentation-for-Diabetic-Retinopathy-Screening-Tests-16-320.jpg)
BEye: Software Implementation and Hardware Acceleration of Retinal Vessel Segmentation for Diabetic Retinopathy Screening Tests
- 1. Retinal vessel segmentation for diabetic retinopathy screening test irene.fidone@mail.polimi.it, lara.cavinato@mail.polimi.it, marco.bacis@mail.polimi.it Irene Fidone, Lara Cavinato, Marco Bacis Da Vinci Conference Room @Stanford OTL May 29th, 2018
- 5. 5 [1] https://nei.nih.gov/health/diabetic/retinopathy Diabetes and Diabetic Retinopathy [1] Approximately 1 in 3 people living with diabetes have some degree of diabetic retinopathy and 1 in 10 will develop a vision impairment form of the disease [2] [2] Yau, Joanne WY, et al. "Global prevalence and major risk factors of diabetic retinopathy." Diabetes care (2012): DC_111909. 53 Type One Diabetes Type Two Diabetes Diabetic Retinopathy 90% Diabetic Retinopathy 60%
- 6. 6 6 Manual segmentationSpecific equipment Qualified staff Screening Test 4
- 7. 7 7Screening Test But sometimes waiting list are too long, and it is possible to wait from 4 months up to more than a year[3] [3] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3553639/pdf/bjgp63-e134.pdf 12-24 months For diabetic patients 6-12 months For diabetic retinopathy patients 3-6 months For diabetic retinopathy patients in the proliferative stage Frequency [months] Guidelines for diabetic retinopathy screening 5
- 8. 8Our Proposal 8 Automated the retinal vessel segmentation for diabetic retinopathy screening Segmentation Algorithm 6
- 9. 99 Software Implementation Hardware acceleration Beye Implementation Segmentation Algorithm 7
- 10. 10Results 108 [4] https://www.isi.uu.nl/Research/Databases/DRIVE/ Accuracy Pixel comparison between manual segmentation and Beye = 92.93% [4] TP + TN Tot Accuracy =
- 11. 11Results 11 [4] https://www.isi.uu.nl/Research/Databases/DRIVE/ Accuracy Execution Time At the same time, our implementation performs 172 times faster respect the state of the art (0.068 s) Pixel comparison between manual segmentation and Beye = 92.93% [4] TP + TN Tot Accuracy = 8
- 12. 12Results 12 [4] https://www.isi.uu.nl/Research/Databases/DRIVE/ Accuracy Execution Time At the same time, our implementation performs 172 times faster respect the state of the art (0.068 s) Pixel comparison between manual segmentation and Beye = 92.93% [4] TP + TN Tot Accuracy = VS. Intel Core i7-6700 CPU Xilinx Zynq-7000 At the same time, our hardware implementation performs 6 times faster respect our software implementation 8
- 14. 10Future Works 1- Convolutional Neural Network for the micro aneurism detection (Non proliferative Phase) 2- Comparison between the segmentation for the proliferation study (Proliferative Phase) 129
- 15. 11 [12] Resnikoff S, Pascolini D, Etya’ale D, et al. Global data on visual impairment in the year 2002. Bulletin of the World Health Organization 2004; 82: 844–51. [13] The Global Economic Cost of Visual Impairment: Summary Report April 2010 It is estimated that Diabetic Retinopathy accounts for 4.8% of the number of cases of blindness worldwide [12] The financial cost of visual impairment worldwide was US$2,954 billion in 2010 [13] Conclusion 1310
- 16. 10Conclusion Thank you for your attention https://www.facebook.com/beye.project/ https://www.slideshare.net/BEye_project https://twitter.com/BEye_NECSTLab irene.fidone@mail.polimi.it It is estimated that Diabetic Retinopathy accounts for 4.8% of the number of cases of blindness worldwide.[12] The financial cost of visual impairment worldwide was US$2,954 billion in 2010[13] 13 May 29th, 2018 Da Vinci Conference Room @Stanford OTL 10