Abstract:Learning pyramidal feature representations is crucial for recognizing object instances at different scales. Feature Pyramid Network (FPN) is the classic architecture to build a feature pyramid with high-level semantics throughout. However, intrinsic defects in feature extraction and fusion inhibit FPN from further aggregating more discriminative features. In this work, we propose Attention Aggregation based Feature Pyramid Network (A 2 -FPN), to improve multi-scale feature learning through attention-guided feature aggregation. In feature extraction, it extracts discriminative features by collecting-distributing multi-level global context features, and mitigates the semantic information loss due to drastically reduced channels. In feature fusion, it aggregates complementary information from adjacent features to generate location-wise reassembly kernels for content-aware sampling, and employs channel-wise reweighting to enhance the semantic consistency before element-wise addition. A 2 -FPN shows consistent gains on different instance segmentation frameworks. By replacing FPN with A 2 -FPN in Mask R-CNN, our model boosts the performance by 2.1% and 1.6% mask AP when using ResNet-50 and ResNet-101 as backbone, respectively. Moreover, A 2 -FPN achieves an improvement of 2.0% and 1.4% mask AP when integrated into the strong baselines such as Cascade Mask R-CNN and Hybrid Task Cascade.

SemSegFPA: Semantic Segmentation Feature Pyramid Network with Attention Module

SAFPN: a Full Semantic Feature Pyramid Network for Object Detection

Deep Dual-Stream Network with Scale Context Selection Attention Module for Semantic Segmentation

PPNet : Pooling Position Attention Network for Semantic Segmentation

Semantic segmentation based on double pyramid network with improved global attention mechanism

Semantic Segmentation Network Based on Adaptive Attention and Deep Fusion Utilizing a Multi-Scale Dilated Convolutional Pyramid

A2-FPN for Semantic Segmentation of Fine-Resolution Remotely Sensed Images

A 2 -FPN for semantic segmentation of fine-resolution remotely sensed images

Point Attention Network for Point Cloud Semantic Segmentation.

FCPFNet: Feature Complementation Network with Pyramid Fusion for Semantic Segmentation

Improving Semantic Segmentation of Aerial Images Using Patch-based Attention

DPANET:Dual Pooling Attention Network for Semantic Segmentation

A<sup>2</sup>-FPN: Attention Aggregation based Feature Pyramid Network for Instance Segmentation

Semantic Image Segmentation Based On Attentions To Intra Scales And Inner Channels

Fully Attentional Network for Semantic Segmentation

Dual-Guided Frequency Prototype Network for Few-Shot Semantic Segmentation

S$^2$-FPN: Scale-ware Strip Attention Guided Feature Pyramid Network for Real-time Semantic Segmentation

MFEAFN: Multi-scale feature enhanced adaptive fusion network for image semantic segmentation

Adaptive multi-scale dual attention network for semantic segmentation

Embedded Attention Network for Semantic Segmentation

A^2-FPN: Attention Aggregation based Feature Pyramid Network for Instance Segmentation