Visualizing the Indicators of Diabetic Retinopathy Learnt by Convolutional Neural Networks

Shikhar Srivastava, Srikanth Prabhu, Sidharth Ramesh, Siddarth Pratapneni, Ashwin Abraham, Sulatha V Bhandary

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This study proposes a novel application of visualizing features learnt by convolutional neural networks with the aim to further the understanding of Diabetic Retinopathy. A convolutional neural network is first trained to recognize and classify fundus images of diabetic and non-diabetic patients. The network is then visualized, using a technique of pixel optimization, to discover the features that the trained network looks for to classify the image. Through this novel application of network visualization, we show that critical features for diabetic retinopathy can be re-discovered, leaving great scope for its application in scarcely explored diseases using minimal resources.

Original languageEnglish
Title of host publication2017 IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017
EditorsN. Krishnan, M. Karthikeyan
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781509066209
DOIs
Publication statusPublished - 05-11-2018
Event8th IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017 - Tamilnadu, India
Duration: 14-12-201716-12-2017

Publication series

Name2017 IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017

Conference

Conference8th IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017
CountryIndia
CityTamilnadu
Period14-12-1716-12-17

Fingerprint

Neural networks
Visualization
Pixels

All Science Journal Classification (ASJC) codes

  • Signal Processing
  • Artificial Intelligence
  • Computational Theory and Mathematics
  • Computer Science Applications

Cite this

Srivastava, S., Prabhu, S., Ramesh, S., Pratapneni, S., Abraham, A., & V Bhandary, S. (2018). Visualizing the Indicators of Diabetic Retinopathy Learnt by Convolutional Neural Networks. In N. Krishnan, & M. Karthikeyan (Eds.), 2017 IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017 [8524578] (2017 IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICCIC.2017.8524578
Srivastava, Shikhar ; Prabhu, Srikanth ; Ramesh, Sidharth ; Pratapneni, Siddarth ; Abraham, Ashwin ; V Bhandary, Sulatha. / Visualizing the Indicators of Diabetic Retinopathy Learnt by Convolutional Neural Networks. 2017 IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017. editor / N. Krishnan ; M. Karthikeyan. Institute of Electrical and Electronics Engineers Inc., 2018. (2017 IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017).
@inproceedings{8021c268701049ddb9d843ab9c40a64f,
title = "Visualizing the Indicators of Diabetic Retinopathy Learnt by Convolutional Neural Networks",
abstract = "This study proposes a novel application of visualizing features learnt by convolutional neural networks with the aim to further the understanding of Diabetic Retinopathy. A convolutional neural network is first trained to recognize and classify fundus images of diabetic and non-diabetic patients. The network is then visualized, using a technique of pixel optimization, to discover the features that the trained network looks for to classify the image. Through this novel application of network visualization, we show that critical features for diabetic retinopathy can be re-discovered, leaving great scope for its application in scarcely explored diseases using minimal resources.",
author = "Shikhar Srivastava and Srikanth Prabhu and Sidharth Ramesh and Siddarth Pratapneni and Ashwin Abraham and {V Bhandary}, Sulatha",
year = "2018",
month = "11",
day = "5",
doi = "10.1109/ICCIC.2017.8524578",
language = "English",
series = "2017 IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
editor = "N. Krishnan and M. Karthikeyan",
booktitle = "2017 IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017",
address = "United States",

}

Srivastava, S, Prabhu, S, Ramesh, S, Pratapneni, S, Abraham, A & V Bhandary, S 2018, Visualizing the Indicators of Diabetic Retinopathy Learnt by Convolutional Neural Networks. in N Krishnan & M Karthikeyan (eds), 2017 IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017., 8524578, 2017 IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017, Institute of Electrical and Electronics Engineers Inc., 8th IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017, Tamilnadu, India, 14-12-17. https://doi.org/10.1109/ICCIC.2017.8524578

Visualizing the Indicators of Diabetic Retinopathy Learnt by Convolutional Neural Networks. / Srivastava, Shikhar; Prabhu, Srikanth; Ramesh, Sidharth; Pratapneni, Siddarth; Abraham, Ashwin; V Bhandary, Sulatha.

2017 IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017. ed. / N. Krishnan; M. Karthikeyan. Institute of Electrical and Electronics Engineers Inc., 2018. 8524578 (2017 IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Visualizing the Indicators of Diabetic Retinopathy Learnt by Convolutional Neural Networks

AU - Srivastava, Shikhar

AU - Prabhu, Srikanth

AU - Ramesh, Sidharth

AU - Pratapneni, Siddarth

AU - Abraham, Ashwin

AU - V Bhandary, Sulatha

PY - 2018/11/5

Y1 - 2018/11/5

N2 - This study proposes a novel application of visualizing features learnt by convolutional neural networks with the aim to further the understanding of Diabetic Retinopathy. A convolutional neural network is first trained to recognize and classify fundus images of diabetic and non-diabetic patients. The network is then visualized, using a technique of pixel optimization, to discover the features that the trained network looks for to classify the image. Through this novel application of network visualization, we show that critical features for diabetic retinopathy can be re-discovered, leaving great scope for its application in scarcely explored diseases using minimal resources.

AB - This study proposes a novel application of visualizing features learnt by convolutional neural networks with the aim to further the understanding of Diabetic Retinopathy. A convolutional neural network is first trained to recognize and classify fundus images of diabetic and non-diabetic patients. The network is then visualized, using a technique of pixel optimization, to discover the features that the trained network looks for to classify the image. Through this novel application of network visualization, we show that critical features for diabetic retinopathy can be re-discovered, leaving great scope for its application in scarcely explored diseases using minimal resources.

UR - http://www.scopus.com/inward/record.url?scp=85057964991&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85057964991&partnerID=8YFLogxK

U2 - 10.1109/ICCIC.2017.8524578

DO - 10.1109/ICCIC.2017.8524578

M3 - Conference contribution

AN - SCOPUS:85057964991

T3 - 2017 IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017

BT - 2017 IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017

A2 - Krishnan, N.

A2 - Karthikeyan, M.

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Srivastava S, Prabhu S, Ramesh S, Pratapneni S, Abraham A, V Bhandary S. Visualizing the Indicators of Diabetic Retinopathy Learnt by Convolutional Neural Networks. In Krishnan N, Karthikeyan M, editors, 2017 IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017. Institute of Electrical and Electronics Engineers Inc. 2018. 8524578. (2017 IEEE International Conference on Computational Intelligence and Computing Research, ICCIC 2017). https://doi.org/10.1109/ICCIC.2017.8524578