Fundus lesion detection based on visual attention model

Baisheng Dai; Wei Bu; Kuanquan Wang; Xiangqian Wu

doi:10.1007/978-981-10-2053-7_34

Fundus lesion detection based on visual attention model

Baisheng Dai
, Wei Bu
, Kuanquan Wang
, Xiangqian Wu^*

^*Corresponding author for this work

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

4 Scopus citations

Abstract

Reliable detection of fundus lesion is important for automated screening of diabetic retinopathy. This paper presents a novel method to detect the fundus lesion in retinal fundus image based on a visual attention model. The proposed method intends to model the visual attention mechanism of ophthalmologists during observing fundus images. That is, the abnormal structures, such as the dark and bright lesions in the image, usually attract the most attention of experts, however, the normal structures, such as optic disc and vessels, have been usually selectively ignored. To measure the visual attention for abnormal and normal areas, the incremental coding length is computed in local and global manner respectively. The final saliency map of fundus lesion is a fusion of attention maps computed for the abnormal and normal areas. Experimental results conducted on the publicly DiaRetDB1 dataset show that the proposed method achieved a sensitivity of 0.71 at a specificity of 0.82 and an AUC of 0.76 for fundus lesion detection, and achieved an accuracy of 100% for normal area (optic disc) detection. The proposed method can assist the ophthalmologists in the inspection of fundus lesion.

Original language	English
Title of host publication	Social Computing - 2nd International Conference of Young Computer Scientists, Engineers and Educators, ICYCSEE 2016, Proceedings
Editors	Wanxiang Che, Hongzhi Wang, Shaoliang Peng, Weipeng Jing, Guanglu Sun, Xianhua Song, Zeguang Lu, Qilong Han, Junyu Lin, Hongtao Song
Publisher	Springer Verlag
Pages	384-394
Number of pages	11
ISBN (Print)	9789811020520
DOIs	https://doi.org/10.1007/978-981-10-2053-7_34
State	Published - 2016
Externally published	Yes
Event	2nd International Conference on Young Computer Scientists, Engineers and Educators, ICYCSEE 2016 - Harbin, China Duration: 20 Aug 2016 → 22 Aug 2016

Publication series

Name	Communications in Computer and Information Science
Volume	623
ISSN (Print)	1865-0929

Conference

Conference	2nd International Conference on Young Computer Scientists, Engineers and Educators, ICYCSEE 2016
Country/Territory	China
City	Harbin
Period	20/08/16 → 22/08/16

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 3 Good Health and Well-being

Keywords

Diabetic retinopathy
Fundus lesion detection
Incremental coding length
Visual attention

Access to Document

10.1007/978-981-10-2053-7_34

Cite this

Dai, B., Bu, W., Wang, K., & Wu, X. (2016). Fundus lesion detection based on visual attention model. In W. Che, H. Wang, S. Peng, W. Jing, G. Sun, X. Song, Z. Lu, Q. Han, J. Lin, & H. Song (Eds.), Social Computing - 2nd International Conference of Young Computer Scientists, Engineers and Educators, ICYCSEE 2016, Proceedings (pp. 384-394). (Communications in Computer and Information Science; Vol. 623). Springer Verlag. https://doi.org/10.1007/978-981-10-2053-7_34

Dai, Baisheng ; Bu, Wei ; Wang, Kuanquan et al. / Fundus lesion detection based on visual attention model. Social Computing - 2nd International Conference of Young Computer Scientists, Engineers and Educators, ICYCSEE 2016, Proceedings. editor / Wanxiang Che ; Hongzhi Wang ; Shaoliang Peng ; Weipeng Jing ; Guanglu Sun ; Xianhua Song ; Zeguang Lu ; Qilong Han ; Junyu Lin ; Hongtao Song. Springer Verlag, 2016. pp. 384-394 (Communications in Computer and Information Science).

@inproceedings{304cba6eeca24667abc8e832032c3261,

title = "Fundus lesion detection based on visual attention model",

abstract = "Reliable detection of fundus lesion is important for automated screening of diabetic retinopathy. This paper presents a novel method to detect the fundus lesion in retinal fundus image based on a visual attention model. The proposed method intends to model the visual attention mechanism of ophthalmologists during observing fundus images. That is, the abnormal structures, such as the dark and bright lesions in the image, usually attract the most attention of experts, however, the normal structures, such as optic disc and vessels, have been usually selectively ignored. To measure the visual attention for abnormal and normal areas, the incremental coding length is computed in local and global manner respectively. The final saliency map of fundus lesion is a fusion of attention maps computed for the abnormal and normal areas. Experimental results conducted on the publicly DiaRetDB1 dataset show that the proposed method achieved a sensitivity of 0.71 at a specificity of 0.82 and an AUC of 0.76 for fundus lesion detection, and achieved an accuracy of 100\% for normal area (optic disc) detection. The proposed method can assist the ophthalmologists in the inspection of fundus lesion.",

keywords = "Diabetic retinopathy, Fundus lesion detection, Incremental coding length, Visual attention",

author = "Baisheng Dai and Wei Bu and Kuanquan Wang and Xiangqian Wu",

note = "Publisher Copyright: {\textcopyright} Springer Science+Business Media Singapore 2016.; 2nd International Conference on Young Computer Scientists, Engineers and Educators, ICYCSEE 2016 ; Conference date: 20-08-2016 Through 22-08-2016",

year = "2016",

doi = "10.1007/978-981-10-2053-7\_34",

language = "英语",

isbn = "9789811020520",

series = "Communications in Computer and Information Science",

publisher = "Springer Verlag",

pages = "384--394",

editor = "Wanxiang Che and Hongzhi Wang and Shaoliang Peng and Weipeng Jing and Guanglu Sun and Xianhua Song and Zeguang Lu and Qilong Han and Junyu Lin and Hongtao Song",

booktitle = "Social Computing - 2nd International Conference of Young Computer Scientists, Engineers and Educators, ICYCSEE 2016, Proceedings",

address = "德国",

}

Dai, B, Bu, W, Wang, K & Wu, X 2016, Fundus lesion detection based on visual attention model. in W Che, H Wang, S Peng, W Jing, G Sun, X Song, Z Lu, Q Han, J Lin & H Song (eds), Social Computing - 2nd International Conference of Young Computer Scientists, Engineers and Educators, ICYCSEE 2016, Proceedings. Communications in Computer and Information Science, vol. 623, Springer Verlag, pp. 384-394, 2nd International Conference on Young Computer Scientists, Engineers and Educators, ICYCSEE 2016, Harbin, China, 20/08/16. https://doi.org/10.1007/978-981-10-2053-7_34

Fundus lesion detection based on visual attention model. / Dai, Baisheng; Bu, Wei; Wang, Kuanquan et al.
Social Computing - 2nd International Conference of Young Computer Scientists, Engineers and Educators, ICYCSEE 2016, Proceedings. ed. / Wanxiang Che; Hongzhi Wang; Shaoliang Peng; Weipeng Jing; Guanglu Sun; Xianhua Song; Zeguang Lu; Qilong Han; Junyu Lin; Hongtao Song. Springer Verlag, 2016. p. 384-394 (Communications in Computer and Information Science; Vol. 623).

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

TY - GEN

T1 - Fundus lesion detection based on visual attention model

AU - Dai, Baisheng

AU - Bu, Wei

AU - Wang, Kuanquan

AU - Wu, Xiangqian

N1 - Publisher Copyright: © Springer Science+Business Media Singapore 2016.

PY - 2016

Y1 - 2016

N2 - Reliable detection of fundus lesion is important for automated screening of diabetic retinopathy. This paper presents a novel method to detect the fundus lesion in retinal fundus image based on a visual attention model. The proposed method intends to model the visual attention mechanism of ophthalmologists during observing fundus images. That is, the abnormal structures, such as the dark and bright lesions in the image, usually attract the most attention of experts, however, the normal structures, such as optic disc and vessels, have been usually selectively ignored. To measure the visual attention for abnormal and normal areas, the incremental coding length is computed in local and global manner respectively. The final saliency map of fundus lesion is a fusion of attention maps computed for the abnormal and normal areas. Experimental results conducted on the publicly DiaRetDB1 dataset show that the proposed method achieved a sensitivity of 0.71 at a specificity of 0.82 and an AUC of 0.76 for fundus lesion detection, and achieved an accuracy of 100% for normal area (optic disc) detection. The proposed method can assist the ophthalmologists in the inspection of fundus lesion.

AB - Reliable detection of fundus lesion is important for automated screening of diabetic retinopathy. This paper presents a novel method to detect the fundus lesion in retinal fundus image based on a visual attention model. The proposed method intends to model the visual attention mechanism of ophthalmologists during observing fundus images. That is, the abnormal structures, such as the dark and bright lesions in the image, usually attract the most attention of experts, however, the normal structures, such as optic disc and vessels, have been usually selectively ignored. To measure the visual attention for abnormal and normal areas, the incremental coding length is computed in local and global manner respectively. The final saliency map of fundus lesion is a fusion of attention maps computed for the abnormal and normal areas. Experimental results conducted on the publicly DiaRetDB1 dataset show that the proposed method achieved a sensitivity of 0.71 at a specificity of 0.82 and an AUC of 0.76 for fundus lesion detection, and achieved an accuracy of 100% for normal area (optic disc) detection. The proposed method can assist the ophthalmologists in the inspection of fundus lesion.

KW - Diabetic retinopathy

KW - Fundus lesion detection

KW - Incremental coding length

KW - Visual attention

UR - https://www.scopus.com/pages/publications/84981169985

U2 - 10.1007/978-981-10-2053-7_34

DO - 10.1007/978-981-10-2053-7_34

M3 - 会议稿件

AN - SCOPUS:84981169985

SN - 9789811020520

T3 - Communications in Computer and Information Science

SP - 384

EP - 394

BT - Social Computing - 2nd International Conference of Young Computer Scientists, Engineers and Educators, ICYCSEE 2016, Proceedings

A2 - Che, Wanxiang

A2 - Wang, Hongzhi

A2 - Peng, Shaoliang

A2 - Jing, Weipeng

A2 - Sun, Guanglu

A2 - Song, Xianhua

A2 - Lu, Zeguang

A2 - Han, Qilong

A2 - Lin, Junyu

A2 - Song, Hongtao

PB - Springer Verlag

T2 - 2nd International Conference on Young Computer Scientists, Engineers and Educators, ICYCSEE 2016

Y2 - 20 August 2016 through 22 August 2016

ER -

Dai B, Bu W, Wang K , Wu X. Fundus lesion detection based on visual attention model. In Che W, Wang H, Peng S, Jing W, Sun G, Song X, Lu Z, Han Q, Lin J, Song H, editors, Social Computing - 2nd International Conference of Young Computer Scientists, Engineers and Educators, ICYCSEE 2016, Proceedings. Springer Verlag. 2016. p. 384-394. (Communications in Computer and Information Science). doi: 10.1007/978-981-10-2053-7_34

Fundus lesion detection based on visual attention model

Abstract

Publication series

Conference

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint

Cite this