guest ::
| Home > Publications Database > Are 2.5D approaches superior to 3D deep networks in whole brain segmentation? > print |
| Home > Publications Database > Are 2.5D approaches superior to 3D deep networks in whole brain segmentation? > print |
| 001 | 280940 | ||
| 005 | 20250918102642.0 | ||
| 037 | _ | _ | |a DZNE-2025-01023 |
| 100 | 1 | _ | |a Roy, Saikat |0 P:(DE-HGF)0 |b 0 |
| 111 | 2 | _ | |a 5th International Conference on Medical Imaging with Deep Learning |g MIDL 2022 |c Zurich |d 2022-07-06 - 2022-07-08 |w Switzerland |
| 245 | _ | _ | |a Are 2.5D approaches superior to 3D deep networks in whole brain segmentation? |
| 260 | _ | _ | |c 2022 |b ML Research Press |
| 295 | 1 | 0 | |a Proceedings of Machine Learning Research |
| 300 | _ | _ | |a 988 - 1004 |
| 336 | 7 | _ | |a CONFERENCE_PAPER |2 ORCID |
| 336 | 7 | _ | |a Conference Paper |0 33 |2 EndNote |
| 336 | 7 | _ | |a INPROCEEDINGS |2 BibTeX |
| 336 | 7 | _ | |a conferenceObject |2 DRIVER |
| 336 | 7 | _ | |a Output Types/Conference Paper |2 DataCite |
| 336 | 7 | _ | |a Contribution to a conference proceedings |b contrib |m contrib |0 PUB:(DE-HGF)8 |s 1758103921_31852 |2 PUB:(DE-HGF) |
| 336 | 7 | _ | |a Contribution to a book |0 PUB:(DE-HGF)7 |2 PUB:(DE-HGF) |m contb |
| 490 | 0 | _ | |v 172 |
| 520 | _ | _ | |a Segmentation of 3D volumes with a large number of labels, small convoluted structures, and lack of contrast between various structural boundaries is a difficult task. While recent methodological advances across many segmentation tasks are dominated by 3D architectures, currently the strongest performing method for whole brain segmentation is FastSurferCNN, a 2.5D approach. To shed light on the nuanced differences between 2.5D and various 3D approaches, we perform a thorough and fair comparison and suggest a spatially-ensembled 3D architecture. Interestingly, we observe training memory intensive 3D segmentation on full-view images does not outperform the 2.5D approach. A shift to training on patches even while evaluating on full-view solves these limitations of both memory and performance limitations at the same time. We demonstrate significant performance improvements over state-of-the-art 3D methods on both Dice Similarity Coefficient and especially average Hausdorff Distance measures across five datasets. Finally, our validation across variations of neurodegenerative disease states and scanner manufacturers, shows we outperform the previously leading 2.5D approach FastSurferCNN demonstrating robust segmentation performance in realistic settings. Our code is available online at github.com/Deep-MI/3d-neuro-seg. |
| 536 | _ | _ | |a 354 - Disease Prevention and Healthy Aging (POF4-354) |0 G:(DE-HGF)POF4-354 |c POF4-354 |f POF IV |x 0 |
| 700 | 1 | _ | |a Kügler, David |0 P:(DE-2719)2814290 |b 1 |u dzne |
| 700 | 1 | _ | |a Reuter, Martin |0 P:(DE-2719)2812134 |b 2 |u dzne |
| 773 | _ | _ | |y 2022 |
| 856 | 4 | _ | |u https://pub.dzne.de/record/280940/files/DZNE-2025-01023_Restricted.pdf |
| 856 | 4 | _ | |u https://pub.dzne.de/record/280940/files/DZNE-2025-01023_Restricted.pdf?subformat=pdfa |x pdfa |
| 909 | C | O | |o oai:pub.dzne.de:280940 |p VDB |
| 910 | 1 | _ | |a Deutsches Zentrum für Neurodegenerative Erkrankungen |0 I:(DE-588)1065079516 |k DZNE |b 1 |6 P:(DE-2719)2814290 |
| 910 | 1 | _ | |a Deutsches Zentrum für Neurodegenerative Erkrankungen |0 I:(DE-588)1065079516 |k DZNE |b 2 |6 P:(DE-2719)2812134 |
| 913 | 1 | _ | |a DE-HGF |b Gesundheit |l Neurodegenerative Diseases |1 G:(DE-HGF)POF4-350 |0 G:(DE-HGF)POF4-354 |3 G:(DE-HGF)POF4 |2 G:(DE-HGF)POF4-300 |4 G:(DE-HGF)POF |v Disease Prevention and Healthy Aging |x 0 |
| 920 | 1 | _ | |0 I:(DE-2719)1040310 |k AG Reuter |l Artificial Intelligence in Medicine |x 0 |
| 980 | _ | _ | |a contrib |
| 980 | _ | _ | |a VDB |
| 980 | _ | _ | |a contb |
| 980 | _ | _ | |a I:(DE-2719)1040310 |
| 980 | _ | _ | |a UNRESTRICTED |
| Library | Collection | CLSMajor | CLSMinor | Language | Author |
|---|