Please use this identifier to cite or link to this item:
http://hdl.handle.net/1942/33812
Full metadata record
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Van Craenendonck, Toon | - |
dc.contributor.author | Elen, Bart | - |
dc.contributor.author | Gerrits, Nele | - |
dc.contributor.author | DE BOEVER, Patrick | - |
dc.date.accessioned | 2021-04-02T11:43:58Z | - |
dc.date.available | 2021-04-02T11:43:58Z | - |
dc.date.issued | 2020 | - |
dc.date.submitted | 2021-03-02T10:26:17Z | - |
dc.identifier.citation | Translational Vision Science & Technology, 9 (2) (Art N° 64) | - |
dc.identifier.uri | http://hdl.handle.net/1942/33812 | - |
dc.description.abstract | Purpose: Heatmapping techniques can support explainability of deep learning (DL) predictions inmedical image analysis. However, individual techniques have been mainly applied in a descriptive way without an objective and systematic evaluation. We investigated comparative performances using diabetic retinopathy lesion detection as a benchmark task. Methods: The Indian Diabetic Retinopathy Image Dataset (IDRiD) publicly available database contains fundus images of diabetes patients with pixel level annotations of diabetic retinopathy (DR) lesions, the ground truth for this study. Three in advance trained DL models (ResNet50, VGG16 or InceptionV3) were used for DR detection in these images. Next, explainability was visualized with each of the 10 most used heatmapping techniques. The quantitative correspondence between the output of a heatmap and the ground truth was evaluated with the Explainability Consistency Score (ECS), a metric between 0 and 1, developed for this comparative task. Results: In case of the overall DR lesions detection, the ECS ranged from 0.21 to 0.51 for all model/heatmapping combinations. The highest score was for VGG16+Grad-CAM (ECS= 0.51; 95% confidence interval [CI]: [0.46; 0.55]). For individual lesions, VGG16+Grad-CAM performed best on hemorrhages and hard exudates. ResNet50+SmoothGrad performed best for soft exudates and ResNet50+Guided Backpropagation performed best for microaneurysms. Conclusions: Our empirical evaluation on the IDRiD database demonstrated that the combination DL model/heatmapping affects explainability when considering common DR lesions. Our approach found considerable disagreement between regions highlighted by heatmaps and expert annotations. | - |
dc.description.sponsorship | Supported by intramural funding from VITO. | - |
dc.language.iso | en | - |
dc.publisher | ASSOC RESEARCH VISION OPHTHALMOLOGY INC | - |
dc.rights | Copyright 2020 The Authors This work is licensed under a Creative Commons Attribution 4.0 International License | - |
dc.subject.other | deep learning | - |
dc.subject.other | heatmap | - |
dc.subject.other | explainability | - |
dc.subject.other | diabetic retinopathy | - |
dc.title | Systematic Comparison of Heatmapping Techniques in Deep Learning in the Context of Diabetic Retinopathy Lesion Detection | - |
dc.type | Journal Contribution | - |
dc.identifier.issue | 2 | - |
dc.identifier.volume | 9 | - |
local.format.pages | 10 | - |
local.bibliographicCitation.jcat | A1 | - |
dc.description.notes | De Boever, P (corresponding author), Univ Antwerp, Dept Biol, Univ Pl 1, B-2610 Antwerp, Belgium. | - |
dc.description.notes | patrick.deboever@uantwerpen.be | - |
dc.description.other | De Boever, P (corresponding author), Univ Antwerp, Dept Biol, Univ Pl 1, B-2610 Antwerp, Belgium. patrick.deboever@uantwerpen.be | - |
local.publisher.place | 12300 TWINBROOK PARKWAY, ROCKVILLE, MD 20852-1606 USA | - |
local.type.refereed | Refereed | - |
local.type.specified | Article | - |
local.bibliographicCitation.artnr | 64 | - |
dc.identifier.doi | 10.1167/tvst.9.2.64 | - |
dc.identifier.isi | WOS:000618898700002 | - |
dc.contributor.orcid | De Boever, Patrick/0000-0002-5197-8215 | - |
local.provider.type | wosris | - |
local.uhasselt.uhpub | yes | - |
local.description.affiliation | [Van Craenendonck, Toon; Elen, Bart; Gerrits, Nele; De Boever, Patrick] VITO NV, Unit Hlth, Mol, Belgium. | - |
local.description.affiliation | [De Boever, Patrick] Hasselt Univ, Hasselt, Belgium. | - |
local.description.affiliation | [De Boever, Patrick] Univ Antwerp, Antwerp, Belgium. | - |
local.uhasselt.international | no | - |
item.fullcitation | Van Craenendonck, Toon; Elen, Bart; Gerrits, Nele & DE BOEVER, Patrick (2020) Systematic Comparison of Heatmapping Techniques in Deep Learning in the Context of Diabetic Retinopathy Lesion Detection. In: Translational Vision Science & Technology, 9 (2) (Art N° 64). | - |
item.validation | ecoom 2022 | - |
item.contributor | Van Craenendonck, Toon | - |
item.contributor | Elen, Bart | - |
item.contributor | Gerrits, Nele | - |
item.contributor | DE BOEVER, Patrick | - |
item.fulltext | With Fulltext | - |
item.accessRights | Open Access | - |
crisitem.journal.issn | 2164-2591 | - |
crisitem.journal.eissn | 2164-2591 | - |
Appears in Collections: | Research publications |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
i2164-2591-9-2-64_1609228844.82789.pdf | Published version | 1.44 MB | Adobe PDF | View/Open |
WEB OF SCIENCETM
Citations
12
checked on May 10, 2024
Page view(s)
36
checked on Sep 5, 2022
Download(s)
12
checked on Sep 5, 2022
Google ScholarTM
Check
Altmetric
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.