{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/3d-anisotropic-hybrid-network-transferring","title":"3D Anisotropic Hybrid Network: Transferring Convolutional Features from 2D Images to 3D Anisotropic Volumes","arxiv_id":"1711.08580","date":"2017-11-23","proceeding":null,"authors":["Si-Qi Liu","Daguang Xu","S. Kevin Zhou","Thomas Mertelmeier","Julia Wicklein","Anna Jerebko","Sasa Grbic","Olivier Pauly","Weidong Cai","Dorin Comaniciu"],"abstract":"While deep convolutional neural networks (CNN) have been successfully applied\nfor 2D image analysis, it is still challenging to apply them to 3D anisotropic\nvolumes, especially when the within-slice resolution is much higher than the\nbetween-slice resolution and when the amount of 3D volumes is relatively small.\nOn one hand, direct learning of CNN with 3D convolution kernels suffers from\nthe lack of data and likely ends up with poor generalization; insufficient GPU\nmemory limits the model size or representational power. On the other hand,\napplying 2D CNN with generalizable features to 2D slices ignores between-slice\ninformation. Coupling 2D network with LSTM to further handle the between-slice\ninformation is not optimal due to the difficulty in LSTM learning. To overcome\nthe above challenges, we propose a 3D Anisotropic Hybrid Network (AH-Net) that\ntransfers convolutional features learned from 2D images to 3D anisotropic\nvolumes. Such a transfer inherits the desired strong generalization capability\nfor within-slice information while naturally exploiting between-slice\ninformation for more effective modelling. The focal loss is further utilized\nfor more effective end-to-end learning. We experiment with the proposed 3D\nAH-Net on two different medical image analysis tasks, namely lesion detection\nfrom a Digital Breast Tomosynthesis volume, and liver and liver tumor\nsegmentation from a Computed Tomography volume and obtain the state-of-the-art\nresults.","url_abs":"http://arxiv.org/abs/1711.08580v2","url_pdf":"http://arxiv.org/pdf/1711.08580v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"3d-anisotropic-hybrid-network-transferring","repo_url":"https://github.com/lsqshr/AH-Net","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":null,"task_name":"GPU"},{"task_slug":"lesion-detection","task_name":"Lesion Detection"},{"task_slug":"medical-image-analysis","task_name":"Medical Image Analysis"},{"task_slug":"tumor-segmentation","task_name":"Tumor Segmentation"}],"methods":[{"method_slug":"3d-convolution","method_name":"3D Convolution"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"focal-loss","method_name":"Focal Loss"},{"method_slug":"lstm","method_name":"LSTM"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"},{"method_slug":"tanh-activation","method_name":"Tanh Activation"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1711.08580","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1711.08580"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/lsqshr/AH-Net","reach":null}],"summary":{"ran_draft_wrong":2},"by_repo_kind":{"listed":{"samples":2,"ran":2,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":2,"samples":[{"code_sha256_prefix":"4487c1cc5125bfe9","entry":"conv3x3x1","repo":"lsqshr/AH-Net","repo_kind":"listed","path":"net3d.py","file_url":"https://github.com/lsqshr/AH-Net/blob/HEAD/net3d.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"4487c1cc5125bfe9"}},{"code_sha256_prefix":"5df53e1b705970c5","entry":"conv3x3x3","repo":"lsqshr/AH-Net","repo_kind":"listed","path":"net3d.py","file_url":"https://github.com/lsqshr/AH-Net/blob/HEAD/net3d.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"5df53e1b705970c5"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}