Benchmarking the robustness of semantic segmentation models with respect to common corruptions
When designing a semantic segmentation model for a real-world application, such as autonomous driving, it is crucial to understand the robustness of the network with respect to a wide range of image corruptions. While there are recent robustness studies for full-image classification, we are the firs...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article (Journal) |
| Language: | English |
| Published: |
2021
|
| In: |
International journal of computer vision
Year: 2021, Volume: 129, Issue: 2, Pages: 462-483 |
| ISSN: | 1573-1405 |
| DOI: | 10.1007/s11263-020-01383-2 |
| Online Access: | Verlag, kostenfrei, Volltext: https://doi.org/10.1007/s11263-020-01383-2
|
| Author Notes: | Christoph Kamann, Carsten Rother |
MARC
| LEADER | 00000caa a2200000 c 4500 | ||
|---|---|---|---|
| 001 | 1776536940 | ||
| 003 | DE-627 | ||
| 005 | 20220208211913.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 211109s2021 xx |||||o 00| ||eng c | ||
| 024 | 7 | |a 10.1007/s11263-020-01383-2 |2 doi | |
| 035 | |a (DE-627)1776536940 | ||
| 035 | |a (DE-599)KXP1776536940 | ||
| 035 | |a (OCoLC)1295678385 | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rda | ||
| 041 | |a eng | ||
| 084 | |a 28 |2 sdnb | ||
| 100 | 1 | |a Kamann, Christoph |d 1993- |e VerfasserIn |0 (DE-588)123637603X |0 (DE-627)1761740598 |4 aut | |
| 245 | 1 | 0 | |a Benchmarking the robustness of semantic segmentation models with respect to common corruptions |c Christoph Kamann, Carsten Rother |
| 264 | 1 | |c 2021 | |
| 300 | |a 22 | ||
| 336 | |a Text |b txt |2 rdacontent | ||
| 337 | |a Computermedien |b c |2 rdamedia | ||
| 338 | |a Online-Ressource |b cr |2 rdacarrier | ||
| 500 | |a Published online: 30 September 2020 | ||
| 500 | |a Gesehen am 09.11.2021 | ||
| 520 | |a When designing a semantic segmentation model for a real-world application, such as autonomous driving, it is crucial to understand the robustness of the network with respect to a wide range of image corruptions. While there are recent robustness studies for full-image classification, we are the first to present an exhaustive study for semantic segmentation, based on many established neural network architectures. We utilize almost 400,000 images generated from the Cityscapes dataset, PASCAL VOC 2012, and ADE20K. Based on the benchmark study, we gain several new insights. Firstly, many networks perform well with respect to real-world image corruptions, such as a realistic PSF blur. Secondly, some architecture properties significantly affect robustness, such as a Dense Prediction Cell, designed to maximize performance on clean data only. Thirdly, the generalization capability of semantic segmentation models depends strongly on the type of image corruption. Models generalize well for image noise and image blur, however, not with respect to digitally corrupted data or weather corruptions. | ||
| 700 | 1 | |a Rother, Carsten |e VerfasserIn |0 (DE-588)1181464692 |0 (DE-627)1662676883 |4 aut | |
| 773 | 0 | 8 | |i Enthalten in |t International journal of computer vision |d Dordrecht [u.a.] : Springer Science + Business Media B.V, 1987 |g 129(2021), 2, Seite 462-483 |h Online-Ressource |w (DE-627)271350083 |w (DE-600)1479903-0 |w (DE-576)102669104 |x 1573-1405 |7 nnas |a Benchmarking the robustness of semantic segmentation models with respect to common corruptions |
| 773 | 1 | 8 | |g volume:129 |g year:2021 |g number:2 |g pages:462-483 |g extent:22 |a Benchmarking the robustness of semantic segmentation models with respect to common corruptions |
| 856 | 4 | 0 | |u https://doi.org/10.1007/s11263-020-01383-2 |x Verlag |x Resolving-System |z kostenfrei |3 Volltext |
| 951 | |a AR | ||
| 992 | |a 20211109 | ||
| 993 | |a Article | ||
| 994 | |a 2021 | ||
| 998 | |g 1181464692 |a Rother, Carsten |m 1181464692:Rother, Carsten |d 700000 |d 708070 |d 700000 |d 728500 |e 700000PR1181464692 |e 708070PR1181464692 |e 700000PR1181464692 |e 728500PR1181464692 |k 0/700000/ |k 1/700000/708070/ |k 0/700000/ |k 1/700000/728500/ |p 2 |y j | ||
| 998 | |g 123637603X |a Kamann, Christoph |m 123637603X:Kamann, Christoph |d 700000 |d 708070 |e 700000PK123637603X |e 708070PK123637603X |k 0/700000/ |k 1/700000/708070/ |p 1 |x j | ||
| 999 | |a KXP-PPN1776536940 |e 4000369172 | ||
| BIB | |a Y | ||
| SER | |a journal | ||
| JSO | |a {"language":["eng"],"recId":"1776536940","note":["Published online: 30 September 2020","Gesehen am 09.11.2021"],"type":{"media":"Online-Ressource","bibl":"article-journal"},"title":[{"title_sort":"Benchmarking the robustness of semantic segmentation models with respect to common corruptions","title":"Benchmarking the robustness of semantic segmentation models with respect to common corruptions"}],"person":[{"given":"Christoph","family":"Kamann","role":"aut","roleDisplay":"VerfasserIn","display":"Kamann, Christoph"},{"roleDisplay":"VerfasserIn","display":"Rother, Carsten","role":"aut","family":"Rother","given":"Carsten"}],"relHost":[{"physDesc":[{"extent":"Online-Ressource"}],"id":{"issn":["1573-1405"],"zdb":["1479903-0"],"eki":["271350083"]},"origin":[{"publisherPlace":"Dordrecht [u.a.] ; Dordrecht [u.a.]","dateIssuedDisp":"1987-","publisher":"Springer Science + Business Media B.V ; Kluwer","dateIssuedKey":"1987"}],"part":{"pages":"462-483","issue":"2","year":"2021","extent":"22","text":"129(2021), 2, Seite 462-483","volume":"129"},"pubHistory":["1.1987 -"],"language":["eng"],"recId":"271350083","type":{"bibl":"periodical","media":"Online-Ressource"},"disp":"Benchmarking the robustness of semantic segmentation models with respect to common corruptionsInternational journal of computer vision","note":["Gesehen am 01.11.05"],"title":[{"title":"International journal of computer vision","title_sort":"International journal of computer vision"}]}],"physDesc":[{"extent":"22 S."}],"id":{"doi":["10.1007/s11263-020-01383-2"],"eki":["1776536940"]},"origin":[{"dateIssuedDisp":"2021","dateIssuedKey":"2021"}],"name":{"displayForm":["Christoph Kamann, Carsten Rother"]}} | ||
| SRT | |a KAMANNCHRIBENCHMARKI2021 | ||