![]() Our extensive experiments validate the effectiveness of our algorithm which outperforms state-of-the-art methods by a significant margin on four widely used few-shot classification benchmarks, namely, miniImageNet, tieredImageNet, Fewshot-CIFAR100 (FC100) and Caltech-UCSD Birds-200-2011 (CUB).Įxperimental results on few-shot learning datasets with ResNet-12 backbone. ![]() Based on the implicit function theorem, the EMD can be inserted as a layer into the network for end-to-end training. To handle k-shot classification, we propose to learn a structured fully connected layer that can directly classify dense image representations with the proposed EMD. To generate the important weights of elements in the EMD formulation, we design a cross-reference mechanism, which can effectively alleviate the adverse impact caused by the cluttered background and large intra-class appearance variations. The EMD generates the optimal matching flows between structural elements that have the minimum matching cost, which is used to calculate the image distance for classification. ![]() We employ the Earth Mover's Distance (EMD) as a metric to compute a structural distance between dense image representations to determine image relevance. In this work, we develop methods for few-shot image classification from a new perspective of optimal matching between image regions. Few-shot learning in contrast attempts to learn with only a few labelled data. Abstractĭeep learning has proved to be very effective in learning with a large amount of labelled data. If you have any question regarding the paper, please send a email to chi007entuedusg. ![]() Differentiable Earth Mover's Distance for Few-Shot Learning},Īuthor=, ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |