Metraipolis › archive › dc_cb108169ff0d
DCM-SAM: Defect-Conditioned Mixture of LoRA Experts for NPU-Deployed AM Defect Segmentation — arXiv:2609.38811
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.38811
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T13:47:20.848Z
- SHA-256 of the archived text
0ac7da667e54bb51ee4289f97583780cece0c99627586b47d9584205a5c6bc35- Stable URL of this page
https://metraipolis.com/documento/dc_cb108169ff0d
How to cite
DCM-SAM: Defect-Conditioned Mixture of LoRA Experts for NPU-Deployed AM Defect Segmentation — arXiv:2609.38811 — arXiv (en), https://arxiv.org/abs/2609.38811 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_cb108169ff0d (Metraipolis archive dc_cb108169ff0d, sha256 0ac7da667e54bb51ee4289f97583780cece0c99627586b47d9584205a5c6bc35)
BibTeX
@misc{dccb108169ff0d,
title = {DCM-SAM: Defect-Conditioned Mixture of LoRA Experts for NPU-Deployed AM Defect Segmentation — arXiv:2609.38811},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.38811},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_cb108169ff0d; sha256 0ac7da667e54bb51ee4289f97583780cece0c99627586b47d9584205a5c6bc35},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.38811. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Metal additive manufacturing parts are inspected by X-ray computed tomography, where labelled data is scarce, the pores and inclusions that matter span a few pixels, and inspection must happen at the machine. We present DCM-SAM, a defect-conditioned adaptive mixture of LoRA experts: one frozen Segment Anything backbone carries a separate Conv-LoRA expert bank and mask decoder per defect class, each trained in its own pass, without prompts, on synthetic slices alone, updating only 4.4% of the parameters. On benchmarks that XCT-SAM reports, DCM-SAM improves on every baseline for both classes from a ViT-B backbone against their ViT-H, and reaches 64.2% pore IoU on real NIST scans having seen no real images during training. Deployment then exposes what adaptation work rarely measures: on a Qualcomm Hexagon NPU, ViT-H and ViT-L compile yet cannot allocate at 1024x1024 image resolution, since activations rather than weights exceed the device ceiling, and quantizing weights does not help. ViT-B alone runs, but the adapted encoder then fails to allocate where the stock one succeeds, until a numerically identical rewrite of the attention lets the complete DCM-SAM run in FP16 at 1024x1024, with no operator falling back to the CPU, masks within 0.01% of pixels of the FP32 reference. Code: this https URL . Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-01 in cs.LG). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2609.38811 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-07T13:18:10.593Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-10-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).