Learning Deep Classifiers Consistent with Fine-Grained Novelty Detection
Learning Deep Classifiers Consistent with Fine-Grained Novelty Detection
复制标题
DOI：
10.1109/cvpr46437.2021.00171
复制
发表时间：
2021-06
期刊：
影响因子：
--
通讯作者：
Jiacheng Cheng;N. Vasconcelos
中科院分区：
文献类型：
--
作者：
Jiacheng Cheng;N. Vasconcelos
研究了细粒度视觉分类（FGVC）中的新颖性检测问题。在卷积神经网络（cnn）的框架内，对基于概率和距离的新颖性检测方法进行了综合理解。研究表明，softmax CNN分类器与新颖性检测不一致，因为它们的学习类条件分布和相关距离度量是不可识别的。然后提出了一个新的正则化约束，即类条件高斯性损失，以消除这种不可辨识性，并强制实现高斯类条件分布。这使得训练新颖性检测一致分类器（ndcc）成为可能，它们对分类和新颖性检测都是最优的。实证评估表明，在小型和大型FGVC数据集上，ndcc都比最先进的方法取得了显著改进。
The problem of novelty detection in fine-grained visual classification (FGVC) is considered. An integrated understanding of the probabilistic and distance-based approaches to novelty detection is developed within the frame-work of convolutional neural networks (CNNs). It is shown that softmax CNN classifiers are inconsistent with novelty detection, because their learned class-conditional distributions and associated distance metrics are unidentifiable. A new regularization constraint, the class-conditional Gaussianity loss, is then proposed to eliminate this unidentifiability, and enforce Gaussian class-conditional distributions. This enables training Novelty Detection Consistent Classifiers (NDCCs) that are jointly optimal for classification and novelty detection. Empirical evaluations show that NDCCs achieve significant improvements over the state-of-the-art on both small- and large-scale FGVC datasets.