Anomaly detection in retinal images using multi-scale deep feature sparse coding

Sourya Dipta Das; Saikat Dutta; Nisarg A. Shah; Dwarikanath Mahapatra; Zongyuan Ge

doi:10.1109/ISBI52829.2022.9761713

Anomaly detection in retinal images using multi-scale deep feature sparse coding

Sourya Dipta Das, Saikat Dutta, Nisarg A. Shah, Dwarikanath Mahapatra, Zongyuan Ge

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

9 Citations (Scopus)

Abstract

Convolutional Neural Network models have successfully detected retinal illness from optical coherence tomography (OCT) and fundus images. These CNN models frequently rely on vast amounts of labeled data for training, difficult to obtain, especially for rare diseases. Furthermore, a deep learning system trained on a data set with only one or a few diseases cannot detect other diseases, limiting the system's practical use in disease identification. We have introduced an unsupervised approach for detecting anomalies in retinal images to overcome this issue. We have proposed a simple, memory efficient, easy to train method which followed a multi-step training technique that incorporated autoencoder training and Multi-Scale Deep Feature Sparse Coding (MDFSC), an extended version of normal sparse coding, to accommodate diverse types of retinal datasets. We achieve relative AUC score improvement of 7.8%, 6.7% and 12.1% over state-of-the-art SPADE on Eye-Q, IDRiD and OCTID datasets respectively.

Original language	English
Title of host publication	ISBI 2022 - Symposium Proceedings
Editors	Suyash P. Awate, Tanveer F. Syeda-Mahmood
Place of Publication	Piscataway NJ USA
Publisher	IEEE, Institute of Electrical and Electronics Engineers
Number of pages	5
ISBN (Electronic)	9781665429238
ISBN (Print)	9781665429245
DOIs	https://doi.org/10.1109/ISBI52829.2022.9761713
Publication status	Published - 2022
Event	IEEE International Symposium on Biomedical Imaging (ISBI) 2022 - Kolkata, India Duration: 28 Mar 2022 → 31 Mar 2022 Conference number: 19th https://ieeexplore.ieee.org/xpl/conhome/9761376/proceeding (Proceedings)

Publication series

Name	Proceedings - International Symposium on Biomedical Imaging
Publisher	IEEE, Institute of Electrical and Electronics Engineers
Volume	2022-March
ISSN (Print)	1945-7928
ISSN (Electronic)	1945-8452

Conference

Conference	IEEE International Symposium on Biomedical Imaging (ISBI) 2022
Abbreviated title	ISBI 2022
Country/Territory	India
City	Kolkata
Period	28/03/22 → 31/03/22
Internet address	https://ieeexplore.ieee.org/xpl/conhome/9761376/proceeding (Proceedings)

Keywords

Anomaly detection
Autoencoder
Deep Multi-scale Feature
Sparse Coding
Transfer Learning

Access to Document

10.1109/ISBI52829.2022.9761713

Cite this

Das, S. D., Dutta, S., Shah, N. A., Mahapatra, D., & Ge, Z. (2022). Anomaly detection in retinal images using multi-scale deep feature sparse coding. In S. P. Awate, & T. F. Syeda-Mahmood (Eds.), ISBI 2022 - Symposium Proceedings (Proceedings - International Symposium on Biomedical Imaging; Vol. 2022-March). IEEE, Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/ISBI52829.2022.9761713

Das, Sourya Dipta ; Dutta, Saikat ; Shah, Nisarg A. et al. / Anomaly detection in retinal images using multi-scale deep feature sparse coding. ISBI 2022 - Symposium Proceedings. editor / Suyash P. Awate ; Tanveer F. Syeda-Mahmood. Piscataway NJ USA : IEEE, Institute of Electrical and Electronics Engineers, 2022. (Proceedings - International Symposium on Biomedical Imaging).

@inproceedings{dc58ce2e69f64a0aa85fe9b9a793fb9e,

title = "Anomaly detection in retinal images using multi-scale deep feature sparse coding",

abstract = "Convolutional Neural Network models have successfully detected retinal illness from optical coherence tomography (OCT) and fundus images. These CNN models frequently rely on vast amounts of labeled data for training, difficult to obtain, especially for rare diseases. Furthermore, a deep learning system trained on a data set with only one or a few diseases cannot detect other diseases, limiting the system's practical use in disease identification. We have introduced an unsupervised approach for detecting anomalies in retinal images to overcome this issue. We have proposed a simple, memory efficient, easy to train method which followed a multi-step training technique that incorporated autoencoder training and Multi-Scale Deep Feature Sparse Coding (MDFSC), an extended version of normal sparse coding, to accommodate diverse types of retinal datasets. We achieve relative AUC score improvement of 7.8%, 6.7% and 12.1% over state-of-the-art SPADE on Eye-Q, IDRiD and OCTID datasets respectively.",

keywords = "Anomaly detection, Autoencoder, Deep Multi-scale Feature, Sparse Coding, Transfer Learning",

author = "Das, {Sourya Dipta} and Saikat Dutta and Shah, {Nisarg A.} and Dwarikanath Mahapatra and Zongyuan Ge",

note = "Publisher Copyright: {\textcopyright} 2022 IEEE.; IEEE International Symposium on Biomedical Imaging (ISBI) 2022, ISBI 2022 ; Conference date: 28-03-2022 Through 31-03-2022",

year = "2022",

doi = "10.1109/ISBI52829.2022.9761713",

language = "English",

isbn = "9781665429245",

series = "Proceedings - International Symposium on Biomedical Imaging",

publisher = "IEEE, Institute of Electrical and Electronics Engineers",

editor = "{P. Awate}, Suyash and {F. Syeda-Mahmood}, Tanveer",

booktitle = "ISBI 2022 - Symposium Proceedings",

address = "United States of America",

url = "https://ieeexplore.ieee.org/xpl/conhome/9761376/proceeding",

}

Das, SD, Dutta, S, Shah, NA, Mahapatra, D & Ge, Z 2022, Anomaly detection in retinal images using multi-scale deep feature sparse coding. in S P. Awate & T F. Syeda-Mahmood (eds), ISBI 2022 - Symposium Proceedings. Proceedings - International Symposium on Biomedical Imaging, vol. 2022-March, IEEE, Institute of Electrical and Electronics Engineers, Piscataway NJ USA, IEEE International Symposium on Biomedical Imaging (ISBI) 2022, Kolkata, India, 28/03/22. https://doi.org/10.1109/ISBI52829.2022.9761713

Anomaly detection in retinal images using multi-scale deep feature sparse coding. / Das, Sourya Dipta; Dutta, Saikat; Shah, Nisarg A. et al.
ISBI 2022 - Symposium Proceedings. ed. / Suyash P. Awate; Tanveer F. Syeda-Mahmood. Piscataway NJ USA: IEEE, Institute of Electrical and Electronics Engineers, 2022. (Proceedings - International Symposium on Biomedical Imaging; Vol. 2022-March).

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

TY - GEN

T1 - Anomaly detection in retinal images using multi-scale deep feature sparse coding

AU - Das, Sourya Dipta

AU - Dutta, Saikat

AU - Shah, Nisarg A.

AU - Mahapatra, Dwarikanath

AU - Ge, Zongyuan

N1 - Conference code: 19th

PY - 2022

Y1 - 2022

N2 - Convolutional Neural Network models have successfully detected retinal illness from optical coherence tomography (OCT) and fundus images. These CNN models frequently rely on vast amounts of labeled data for training, difficult to obtain, especially for rare diseases. Furthermore, a deep learning system trained on a data set with only one or a few diseases cannot detect other diseases, limiting the system's practical use in disease identification. We have introduced an unsupervised approach for detecting anomalies in retinal images to overcome this issue. We have proposed a simple, memory efficient, easy to train method which followed a multi-step training technique that incorporated autoencoder training and Multi-Scale Deep Feature Sparse Coding (MDFSC), an extended version of normal sparse coding, to accommodate diverse types of retinal datasets. We achieve relative AUC score improvement of 7.8%, 6.7% and 12.1% over state-of-the-art SPADE on Eye-Q, IDRiD and OCTID datasets respectively.

AB - Convolutional Neural Network models have successfully detected retinal illness from optical coherence tomography (OCT) and fundus images. These CNN models frequently rely on vast amounts of labeled data for training, difficult to obtain, especially for rare diseases. Furthermore, a deep learning system trained on a data set with only one or a few diseases cannot detect other diseases, limiting the system's practical use in disease identification. We have introduced an unsupervised approach for detecting anomalies in retinal images to overcome this issue. We have proposed a simple, memory efficient, easy to train method which followed a multi-step training technique that incorporated autoencoder training and Multi-Scale Deep Feature Sparse Coding (MDFSC), an extended version of normal sparse coding, to accommodate diverse types of retinal datasets. We achieve relative AUC score improvement of 7.8%, 6.7% and 12.1% over state-of-the-art SPADE on Eye-Q, IDRiD and OCTID datasets respectively.

KW - Anomaly detection

KW - Autoencoder

KW - Deep Multi-scale Feature

KW - Sparse Coding

KW - Transfer Learning

UR - http://www.scopus.com/inward/record.url?scp=85129614026&partnerID=8YFLogxK

U2 - 10.1109/ISBI52829.2022.9761713

DO - 10.1109/ISBI52829.2022.9761713

M3 - Conference Paper

AN - SCOPUS:85129614026

SN - 9781665429245

T3 - Proceedings - International Symposium on Biomedical Imaging

BT - ISBI 2022 - Symposium Proceedings

A2 - P. Awate, Suyash

A2 - F. Syeda-Mahmood, Tanveer

PB - IEEE, Institute of Electrical and Electronics Engineers

CY - Piscataway NJ USA

T2 - IEEE International Symposium on Biomedical Imaging (ISBI) 2022

Y2 - 28 March 2022 through 31 March 2022

ER -

Das SD, Dutta S, Shah NA, Mahapatra D, Ge Z. Anomaly detection in retinal images using multi-scale deep feature sparse coding. In P. Awate S, F. Syeda-Mahmood T, editors, ISBI 2022 - Symposium Proceedings. Piscataway NJ USA: IEEE, Institute of Electrical and Electronics Engineers. 2022. (Proceedings - International Symposium on Biomedical Imaging). doi: 10.1109/ISBI52829.2022.9761713

Anomaly detection in retinal images using multi-scale deep feature sparse coding

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Cite this