DCM-SAM：面向NPU部署的增材制造缺陷分割的缺陷条件LoRA专家混合模型
DCM-SAM: Defect-Conditioned Mixture of LoRA Experts for NPU-Deployed AM Defect Segmentation
- West Virginia University（西弗吉尼亚大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
提出DCM-SAM，一种缺陷条件LoRA专家混合模型，在合成数据上训练，仅更新4.4%参数，在XCT-SAM基准和真实NIST扫描上超越基线，并通过注意力重写实现NPU上FP16高效部署。
AI中文摘要：
金属增材制造零件通过X射线计算机断层扫描进行检测，其中标注数据稀缺，关键的孔隙和夹杂物仅跨越几个像素，且检测必须在机器上进行。我们提出DCM-SAM，一种缺陷条件的自适应LoRA专家混合模型：一个冻结的Segment Anything骨干网络为每个缺陷类别携带独立的Conv-LoRA专家库和掩码解码器，每个类别在各自的训练过程中，仅使用合成切片，无需提示，仅更新4.4%的参数。在XCT-SAM报告的基准上，DCM-SAM在ViT-B骨干网络下对两个类别均优于所有基线（对比其ViT-H），并在未见任何真实图像的情况下，在真实NIST扫描上达到64.2%的孔隙IoU。部署随后暴露了适应工作很少衡量的内容：在Qualcomm Hexagon NPU上，ViT-H和ViT-L可编译但无法在1024x1024图像分辨率下分配内存，因为激活而非权重超出设备上限，且量化权重无济于事。仅ViT-B可运行，但适应后的编码器在原始编码器成功之处分配失败，直到对注意力进行数值上等效的重写，使完整的DCM-SAM以FP16在1024x1024下运行，无算子回退至CPU，掩码与FP32参考的像素差异在0.01%以内。代码：此https URL。
英文摘要：
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: https://github.com/MushfiqShovon/DCM-SAM.
