Multi-level Second-order Few-shot Learning
arXiv:2201.05916 · doi:10.1109/TMM.2022.3142955
Abstract
We propose a Multi-level Second-order (MlSo) few-shot learning network for supervised or unsupervised few-shot image classification and few-shot action recognition. We leverage so-called power-normalized second-order base learner streams combined with features that express multiple levels of visual abstraction, and we use self-supervised discriminating mechanisms. As Second-order Pooling (SoP) is popular in image recognition, we employ its basic element-wise variant in our pipeline. The goal of multi-level feature design is to extract feature representations at different layer-wise levels of CNN, realizing several levels of visual abstraction to achieve robust few-shot learning. As SoP can handle convolutional feature maps of varying spatial sizes, we also introduce image inputs at multiple spatial scales into MlSo. To exploit the discriminative information from multi-level and multi-scale features, we develop a Feature Matching (FM) module that reweights their respective branches. We also introduce a self-supervised step, which is a discriminator of the spatial level and the scale of abstraction. Our pipeline is trained in an end-to-end manner. With a simple architecture, we demonstrate respectable results on standard datasets such as Omniglot, mini-ImageNet, tiered-ImageNet, Open MIC, fine-grained datasets such as CUB Birds, Stanford Dogs and Cars, and action recognition datasets such as HMDB51, UCF101, and mini-MIT.
arXiv commentIEEE Transactions on Multimedia
References in corpus (19)
- Semi-Supervised Classification with Graph Convolutional Networks
- A Simple Framework for Contrastive Learning of Visual Representations
- Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
- Prototypical Networks for Few-shot Learning
- UCF101: A Dataset of 101 Human Actions Classes From Videos in The Wild
- Simplifying Graph Convolutional Networks
- Meta-SGD: Learning to Learn Quickly for Few-Shot Learning
- Meta-Learning for Semi-Supervised Few-Shot Classification
- Learning to Propagate Labels: Transductive Propagation Network for Few-shot Learning
- Deep Clustering for Unsupervised Learning of Visual Features
- Predict then Propagate: Graph Neural Networks meet Personalized PageRank
- SimpleShot: Revisiting Nearest-Neighbor Classification for Few-Shot Learning
- Low-Rank Pairwise Alignment Bilinear Network For Few-Shot Fine-Grained Image Classification
- Fast Context Adaptation via Meta-Learning
- Unsupervised Learning via Meta-Learning
- Power Normalizations in Fine-grained Image, Few-shot Image and Graph Classification
- Tensor Representations for Action Recognition
- Relation Embedding for Personalised POI Recommendation
- FLAT: Few-Shot Learning via Autoencoding Transformation Regularizers