File Download
There are no files associated with this item.
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.1007/978-3-030-59719-1_53
- Scopus: eid_2-s2.0-85092777475
Supplementary
-
Citations:
- Scopus: 0
- Appears in Collections:
Conference Paper: Revisiting 3D Context Modeling with Supervised Pre-training for Universal Lesion Detection in CT Slices
Title | Revisiting 3D Context Modeling with Supervised Pre-training for Universal Lesion Detection in CT Slices |
---|---|
Authors | |
Keywords | Lesion detection 3D context modeling 3D network pre-training |
Issue Date | 2020 |
Publisher | Springer. |
Citation | The 23rd International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI 2020), Virtual Conference, Lima, Peru, 4-8 October 2020. In MICCAI 2020 Proceedings, Part 4, p. 542-551 How to Cite? |
Abstract | Universal lesion detection from computed tomography (CT) slices is important for comprehensive disease screening. Since each lesion can locate in multiple adjacent slices, 3D context modeling is of great significance for developing automated lesion detection algorithms. In this work, we propose a Modified Pseudo-3D Feature Pyramid Network (MP3D FPN) that leverages depthwise separable convolutional filters and a group transform module (GTM) to efficiently extract 3D context enhanced 2D features for universal lesion detection in CT slices. To facilitate faster convergence, a novel 3D network pre-training method is derived using solely large-scale 2D object detection dataset in the natural image domain. We demonstrate that with the novel pre-training method, the proposed MP3D FPN achieves state-of-the-art detection performance on the DeepLesion dataset (3.48% absolute improvement in the sensitivity of FPs@0.5), significantly surpassing the baseline method by up to 6.06% (in MAP@0.5) which adopts 2D convolution for 3D context modeling. Moreover, the proposed 3D pre-trained weights can potentially be used to boost the performance of other 3D medical image analysis tasks. |
Persistent Identifier | http://hdl.handle.net/10722/301186 |
ISBN | |
Series/Report no. | Lecture Notes in Computer Science (LNCS) ; v. 12264 |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Zhang, S | - |
dc.contributor.author | Xu, J | - |
dc.contributor.author | Chen, Y-C | - |
dc.contributor.author | Ma, J | - |
dc.contributor.author | Li, Z | - |
dc.contributor.author | Wang, Y | - |
dc.contributor.author | Yu, Y | - |
dc.date.accessioned | 2021-07-27T08:07:23Z | - |
dc.date.available | 2021-07-27T08:07:23Z | - |
dc.date.issued | 2020 | - |
dc.identifier.citation | The 23rd International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI 2020), Virtual Conference, Lima, Peru, 4-8 October 2020. In MICCAI 2020 Proceedings, Part 4, p. 542-551 | - |
dc.identifier.isbn | 9783030597184 | - |
dc.identifier.uri | http://hdl.handle.net/10722/301186 | - |
dc.description.abstract | Universal lesion detection from computed tomography (CT) slices is important for comprehensive disease screening. Since each lesion can locate in multiple adjacent slices, 3D context modeling is of great significance for developing automated lesion detection algorithms. In this work, we propose a Modified Pseudo-3D Feature Pyramid Network (MP3D FPN) that leverages depthwise separable convolutional filters and a group transform module (GTM) to efficiently extract 3D context enhanced 2D features for universal lesion detection in CT slices. To facilitate faster convergence, a novel 3D network pre-training method is derived using solely large-scale 2D object detection dataset in the natural image domain. We demonstrate that with the novel pre-training method, the proposed MP3D FPN achieves state-of-the-art detection performance on the DeepLesion dataset (3.48% absolute improvement in the sensitivity of FPs@0.5), significantly surpassing the baseline method by up to 6.06% (in MAP@0.5) which adopts 2D convolution for 3D context modeling. Moreover, the proposed 3D pre-trained weights can potentially be used to boost the performance of other 3D medical image analysis tasks. | - |
dc.language | eng | - |
dc.publisher | Springer. | - |
dc.relation.ispartof | International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) | - |
dc.relation.ispartofseries | Lecture Notes in Computer Science (LNCS) ; v. 12264 | - |
dc.subject | Lesion detection | - |
dc.subject | 3D context modeling | - |
dc.subject | 3D network pre-training | - |
dc.title | Revisiting 3D Context Modeling with Supervised Pre-training for Universal Lesion Detection in CT Slices | - |
dc.type | Conference_Paper | - |
dc.identifier.email | Yu, Y: yzyu@cs.hku.hk | - |
dc.identifier.authority | Yu, Y=rp01415 | - |
dc.description.nature | link_to_subscribed_fulltext | - |
dc.identifier.doi | 10.1007/978-3-030-59719-1_53 | - |
dc.identifier.scopus | eid_2-s2.0-85092777475 | - |
dc.identifier.hkuros | 323537 | - |
dc.identifier.volume | Part 4 | - |
dc.identifier.spage | 542 | - |
dc.identifier.epage | 551 | - |
dc.publisher.place | Cham | - |