图像分类中面向分布外鲁棒性的贝叶斯方法
A Bayesian Approach to OOD Robustness in Image Classification
AI总结：
本文提出无监督生成过渡（UGT）方法，通过贝叶斯框架和vMF核过渡字典扩展CompNets，在无目标域标注的情况下提升图像分类对分布外干扰和遮挡的鲁棒性，并在多个OOD基准上超越现有方法。
摘要：
计算机视觉中一个重要且尚未解决的问题是确保算法对图像域的变化具有鲁棒性。我们在能够获取目标域图像但没有标注的场景下解决该问题。受OOD-CV基准测试中遇到的真实世界分布外（OOD）干扰和遮挡挑战的启发，我们提出了一种新颖的面向物体分类分布外鲁棒性的贝叶斯方法。我们的工作扩展了组合神经网络（CompNets），该网络已被证明对遮挡具有鲁棒性，但在分布外数据上测试时性能严重下降。我们利用的事实是，CompNets包含一个定义在由von Mises-Fisher（vMF）核表示的特征向量上的生成头，这些核大致对应于物体部件，并且可以在无监督条件下学习。我们观察到，一些vMF核在不同域之间是相似的，而另一些则不是。这使我们能够学习一个介于源域和目标域之间的vMF核过渡字典，并使用源域上的标注在该字典上训练生成模型，随后进行迭代细化。该方法被称为无监督生成过渡（UGT），即使在存在遮挡的情况下，在OOD场景中也表现非常好。UGT在不同的OOD基准上进行了评估，包括OOD-CV数据集、多个流行数据集（例如ImageNet-C [9]）、人工图像损坏（包括添加遮挡物）以及合成到真实的域迁移，并且在所有场景中表现良好，优于最先进的替代方法（例如，在Occluded OOD-CV数据集上top-1准确率提升高达10%）。
英文摘要：
An important and unsolved problem in computer vision is to ensure that the algorithms are robust to changes in image domains. We address this problem in the scenario where we have access to images from the target domains but no annotations. Motivated by the challenges of the OOD-CV benchmark where we encounter real world Out-of-Domain (OOD) nuisances and occlusion, we introduce a novel Bayesian approach to OOD robustness for object classification. Our work extends Compositional Neural Networks (CompNets), which have been shown to be robust to occlusion but degrade badly when tested on OOD data. We exploit the fact that CompNets contain a generative head defined over feature vectors represented by von Mises-Fisher (vMF) kernels, which correspond roughly to object parts, and can be learned without supervision. We obverse that some vMF kernels are similar between different domains, while others are not. This enables us to learn a transitional dictionary of vMF kernels that are intermediate between the source and target domains and train the generative model on this dictionary using the annotations on the source domain, followed by iterative refinement. This approach, termed Unsupervised Generative Transition (UGT), performs very well in OOD scenarios even when occlusion is present. UGT is evaluated on different OOD benchmarks including the OOD-CV dataset, several popular datasets (e.g., ImageNet-C [9]), artificial image corruptions (including adding occluders), and synthetic-to-real domain transfer, and does well in all scenarios outperforming SOTA alternatives (e.g. up to 10% top-1 accuracy on Occluded OOD-CV dataset).
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
