InvAD：基于反演的无重建扩散模型异常检测
InvAD: Inversion-based Reconstruction-Free Anomaly Detection with Diffusion Models
把扩散模型的 DDIM 反演当作无需重建的异常打分器：只需少量反演步把测试图加噪到潜变量，再用已知先验分布度量偏离度，可直接替换重建式 AD 中昂贵的多步去噪流程，兼顾精度与约两倍推理加速。
原文依据 we model AD under a reconstruction-free formulation that directly infers the final latent variable corresponding to the input image via DDIM inversion
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论尽管近期基于扩散建模的重建式异常检测（AD）方法取得了显著成功，但它们仍涉及细粒度的噪声强度调参和计算代价高昂的多步去噪，导致保真度与效率之间的根本性矛盾。本文提出 InvAD，一种新颖的基于反演的异常检测方法，通过在潜变量空间中加噪完成检测，从而规避了显式重建。重要的是，我们认为先前重建式方法的局限源于盛行的「在 RGB 空间中通过去噪进行检测」范式。为此，我们在无重建的表述下对 AD 建模：通过 DDIM 反演直接推断与输入图像对应的最终潜变量，然后基于已知先验分布度量偏离程度以进行异常打分。具体而言，在使用欧拉法近似原始概率流 ODE 时，我们仅用少量反演步即可对干净图像加噪，从而提升推理效率。由于所加噪声自适应地源自学习到的扩散模型，干净测试图像的原始特征仍可被利用以取得高检测精度。我们在无监督统一设定下于四个广泛使用的 AD 基准上进行了大量实验与详细分析，证明了模型的有效性，取得了最优的 AD 性能，并在不进行扩散蒸馏的情况下实现约两倍的推理加速。
英文原文摘要 1529 字符
Despite the remarkable success, recent reconstruction-based anomaly detection (AD) methods via diffusion modeling still involve fine-grained noise-strength tuning and computationally expensive multi-step denoising, leading to a fundamental tension between fidelity and efficiency. In this paper, we propose InvAD, a novel inversion-based anomaly detection approach for detection via noising in latent space, which circumvents explicit reconstruction. Importantly, we contend that the limitations of prior reconstruction-based methods originate from the prevailing "detection via denoising in RGB space" paradigm. To address this, we model AD under a reconstruction-free formulation that directly infers the final latent variable corresponding to the input image via DDIM inversion and then measures the deviation based on the known prior distribution for anomaly scoring. Specifically, when approximating the original probability flow ODE using the Euler method, we use only a few inversion steps to noise the clean image and improve inference efficiency. As the added noise is adaptively derived from the learned diffusion model, the original features of the clean test image can still be leveraged to yield high detection accuracy. We perform extensive experiments and detailed analysis across four widely used AD benchmarks under the unsupervised unified setting to demonstrate the effectiveness of our model, achieving state-of-the-art AD performance and about a twofold inference-time speedup without diffusion distillation.