Abstract:Few-shot object detection (FSOD) aims at designing models that can accurately detect targets of novel classes in a scarce data regime. Existing research has improved detection performance with meta-learning-based models. However, existing methods continue to exhibit certain imperfections: (1) Only the interacting global features of query and support images lead to ignoring local critical features in the imprecise localization of objects from new categories. (2) Convolutional neural networks (CNNs) encounter difficulty in learning diverse pose features from exceedingly limited labeled samples of unseen classes. (3) Local context information is not fully utilized in a global attention mechanism, which means the attention modules need to be improved. As a result, the detection performance of novel-class objects is compromised. To overcome these challenges, a few-shot object detection network is proposed with a local feature enhancement module and an intrinsic feature transformation module. In this paper, a local feature enhancement module (LFEM) is designed to raise the importance of intrinsic features of the novel-class samples. In addition, an Intrinsic Feature Transform Module (IFTM) is explored to enhance the feature representation of novel-class samples, which enriches the feature space of novel classes. Finally, a more effective cross-attention module, called Global Cross-Attention Network (GCAN), which fully aggregates local and global context information between query and support images, is proposed in this paper. The crucial features of novel-class objects are extracted effectively by our model before the feature fusion between query images and support images. Our proposed method increases, on average, the detection performance by 0.93 (nAP) in comparison with previous models on the PASCAL VOC FSOD benchmark dataset. Extensive experiments demonstrate the effectiveness of our modules under various experimental settings.

Few-shot Classification with Intra-Class Co-Salient Learning and Holistic Metric

Multi-Level Correlation Network For Few-Shot Image Classification

A Two-Stream Network with Image-to-Class Deep Metric for Few-Shot Classification.

Few-Shot Object Detection with Local Feature Enhancement and Feature Interrelation

Coarse-to-fine Few-Shot Classification with Deep Metric Learning

Few-shot Learning for Domain-specific Fine-grained Image Classification

Task-Related Saliency for Few-Shot Image Classification

Few-Shot Fine-Grained Image Classification via Multi-Frequency Neighborhood and Double-Cross Modulation

Class Feature Sub-space for Few-Shot Classification

Class-based Core Feature Extraction Network for Few-shot Classification

Channel-spatial attention network for fewshot classification

Object-aware Long-short-range Spatial Alignment for Few-Shot Fine-Grained Image Classification

Cross-Domain Few-Shot Hyperspectral Image Classification With Cross-Modal Alignment and Supervised Contrastive Learning

Local Mutual Metric Network for Few-Shot Image Classification

Hyperspectral Image Few-Shot Classification Network with Brownian Distance Covariance.

Sampling-invariant fully metric learning for few-shot object detection

MHA-WoML: Multi-head Attention and Wasserstein-OT for Few-Shot Learning

Discriminant Space Metric Network for Few-Shot Image Classification

Spatial-Spectral-Semantic Cross-Domain Few-Shot Learning for Hyperspectral Image Classification

Learning to focus: cascaded feature matching network for few-shot image recognition

Few-shot Image Classification Via Hybrid Representation