283020 20260103103701.0 10.1016/j.inffus.2025.103823 doi 1566-2535 ISSN 1872-6305 ISSN DZNE-2025-01432 620 Zhao, Wenzhao 0000-0001-5150-3781 0 Efficient 3D affinely equivariant CNNs with adaptive fusion of augmented spherical Fourier–Bessel bases Amsterdam [u.a.] 2026 Elsevier Science article DRIVER Output Types/Journal article DataCite Journal Article journal journal PUB:(DE-HGF)16 1767358875_32512 PUB:(DE-HGF) ARTICLE BibTeX JOURNAL_ARTICLE ORCID Journal Article 0 EndNote The code is available at https://github.com/ZhaoWenzhao/WMCSFB. Filter-decomposition-based group equivariant convolutional neural networks (CNNs) have shown promising stability and data efficiency for 3D image feature extraction. However, these networks, which rely on parameter sharing and discrete transformation groups, often underperform in modern deep neural network architectures for processing volumetric images with dense 3D textures, such as the common 3D medical images. To address these limitations, this paper presents an efficient non-parameter-sharing continuous 3D affine group equivariant neural network for volumetric images. This network uses an adaptive aggregation of Monte Carlo augmented spherical Fourier–Bessel filter bases to improve the efficiency and flexibility of 3D group equivariant CNNs for volumetric data. Unlike existing methods that focus only on angular orthogonality in filter bases, the introduced spherical Bessel Fourier filter base incorporates both angular and radial orthogonality to improve feature extraction. Experiments on four medical image segmentation datasets and two seismic datasets show that the proposed methods achieve better affine group equivariance and superior segmentation accuracy than existing 3D group equivariant convolutional neural network layers, significantly improving the training stability and data efficiency of conventional CNN layers (at 0.05 significance level). 353 - Clinical and Health Care Research (POF4-353) G:(DE-HGF)POF4-353 POF4-353 POF IV 0 Dataset connected to CrossRef, Journals: pub.dzne.de Albert, Steffen 1 Wichtmann, Barbara D. P:(DE-2719)9003380 2 dzne Schmitt, Angelika P:(DE-2719)2810830 3 dzne Attenberger, Ulrike 4 Zöllner, Frank G. 5 Hesser, Jürgen 6 10.1016/j.inffus.2025.103823 Vol. 127, p. 103823 - PERI:(DE-600)2025632-2 103823 Information fusion 127 2026 1566-2535 http://pub.dzne.de/record/283020/files/DZNE-2025-1432_Restricted.pdf http://pub.dzne.de/record/283020/files/DZNE-2025-1432_Restricted.pdf?subformat=pdfa pdfa oai:pub.dzne.de:283020 VDB Deutsches Zentrum für Neurodegenerative Erkrankungen I:(DE-588)1065079516 DZNE 2 P:(DE-2719)9003380 External Institute I:(DE-HGF)0 Extern 3 P:(DE-2719)2810830 DE-HGF Gesundheit Neurodegenerative Diseases G:(DE-HGF)POF4-350 G:(DE-HGF)POF4-353 G:(DE-HGF)POF4 G:(DE-HGF)POF4-300 G:(DE-HGF)POF Clinical and Health Care Research 0 2026 DBCoverage StatID:(DE-HGF)0200 StatID SCOPUS 2025-01-02 DBCoverage StatID:(DE-HGF)0300 StatID Medline 2025-01-02 DBCoverage StatID:(DE-HGF)0199 StatID Clarivate Analytics Master Journal List 2025-01-02 DBCoverage StatID:(DE-HGF)1160 StatID Current Contents - Engineering, Computing and Technology 2025-01-02 DBCoverage StatID:(DE-HGF)0160 StatID Essential Science Indicators 2025-01-02 WoS StatID:(DE-HGF)0113 StatID Science Citation Index Expanded 2025-01-02 DBCoverage StatID:(DE-HGF)0150 StatID Web of Science Core Collection 2025-01-02 DBCoverage StatID:(DE-HGF)0600 StatID Ebsco Academic Search 2025-01-02 Peer Review StatID:(DE-HGF)0030 StatID ASC 2025-01-02 I:(DE-2719)5000075 AG Radbruch Clinical Neuroimaging 0 journal VDB I:(DE-2719)5000075 UNRESTRICTED