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.

Meta-RetinaNet for Few-shot Object Detection

Meta-SSD: Towards Fast Adaptation for Few-Shot Object Detection with Meta-Learning

Meta-Learning-Based Incremental Few-Shot Object Detection

Cascade Meta-RCNN for Few-shot Object Detection

An Adaptive Detector for Few Shot Object Detection.

Meta-Generating Deep Attentive Metric for Few-Shot Classification

Meta-YOLO: Meta-Learning for Few-Shot Traffic Sign Detection via Decoupling Dependencies

Few-Shot Directed Meta-Learning for Image Classification

Few-Shot Object Detection in Remote Sensing Images via Data Clearing and Stationary Meta-Learning

MM-FSOD: Meta and metric integrated few-shot object detection

Feature Reconstruction and Metric Based Network for Few-Shot Object Detection

Meta-Learning Adversarial Domain Adaptation Network for Few-Shot Text Classification.

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

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

Meta Faster R-CNN: Towards Accurate Few-Shot Object Detection with Attentive Feature Alignment

Meta-Adapter: An Online Few-shot Learner for Vision-Language Model

MPF-Net: multi-projection filtering network for few-shot object detection

Efficient Meta-Learning Enabled Lightweight Multiscale Few-Shot Object Detection in Remote Sensing Images

Meta-tuning Loss Functions and Data Augmentation for Few-shot Object Detection

Meta-Det3D: Learn to Learn Few-Shot 3D Object Detection

Dynamic Relevance Learning for Few-Shot Object Detection