Fine-Grained Few-Shot Classification with Feature Map Reconstruction Networks

By Davis Wertheimer, Luming Tang, Bharath Hariharan. arXiv.org, 2020

315 citations recorded by the source corpus.

In this paper we reformulate few-shot classification as a reconstruction problem in latent space. The ability of the network to reconstruct a query feature map from support features of a given class predicts membership of the query in that class. We introduce a novel mechanism for few-shot classification by regressing directly from support features to query features in closed form, without introducing any new modules or large-scale learnable parameters. The resulting Feature Map Reconstruction Networks are both more performant and computationally efficient than previous approaches. We demonstrate consistent and significant accuracy gains on four fine-grained benchmarks with varying neural architectures. Our model is also competitive on the non-fine-grained mini-ImageNet benchmark with minimal bells and whistles.

View Paper: https://www.semanticscholar.org/paper/f4c58e6ebd138bd5156f371b310c84fd0c324bbd

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

1. Junyi Cao
2. Chao Ma
3. Taiping Yao
4. Shouhong Ding
5. Xiaokang Yang
6. Shen Chen
7. Yisheng Song
8. Ting-Yuan Wang

Citations in the corpus

• A Comprehensive Survey of Few-shot Learning: Evolution, Applications, Challenges, and Opportunities (2022, background): “Based on this, FRN [84] explores the potential space for few-shot image classification, using ridge regression to reconstruct and normalize the feature map without adding new learning parameters.”
• End-to-End Reconstruction-Classification Learning for Face Forgery Detection (2022, background): “Reconstruction learning has been wildly used for representation learning in unsupervised settings [16,26,32,49,51,52].”