Global Aggregation then Local Distribution in Fully Convolutional Networks

Abstract

It has been widely proven that modelling long-range dependencies in fully convolutional networks (FCNs) via global aggregation modules is critical for complex scene understanding tasks such as semantic segmentation and object detection. However, global aggregation is often dominated by features of large patterns and tends to oversmooth regions that contain small patterns (e.g., boundaries and small objects). To resolve this problem, we propose to first use \emph{Global Aggregation} and then \emph{Local Distribution}, which is called GALD, where long-range dependencies are more confidently used inside large pattern regions and vice versa. The size of each pattern at each position is estimated in the network as a per-channel mask map. GALD is end-to-end trainable and can be easily plugged into existing FCNs with various global aggregation modules for a wide range of vision tasks, and consistently improves the performance of state-of-the-art object detection and instance segmentation approaches. In particular, GALD used in semantic segmentation achieves new state-of-the-art performance on Cityscapes test set with mIoU 83.3\%. Code is available at: \url{https://github.com/lxtGH/GALD-Net}

arXiv commentaccepted at BMVC 2019

References in corpus (13)

• Rethinking Atrous Convolution for Semantic Image Segmentation
• Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation
• R-FCN: Object Detection via Region-based Fully Convolutional Networks
• DSSD : Deconvolutional Single Shot Detector
• ParseNet: Looking Wider to See Better
• The Cityscapes Dataset for Semantic Urban Scene Understanding
• Understanding the Effective Receptive Field in Deep Convolutional Neural Networks
• OCNet: Object Context Network for Scene Parsing
• Searching for Efficient Multi-Scale Architectures for Dense Image Prediction
• Auto-DeepLab: Hierarchical Neural Architecture Search for Semantic Image Segmentation
• Context Encoding for Semantic Segmentation
• Learning a Discriminative Feature Network for Semantic Segmentation
• High-performance Semantic Segmentation Using Very Deep Fully Convolutional Networks

Cited by in corpus (14)

• A Survey on Visual Transformer
• Segmentation Transformer: Object-Contextual Representations for Semantic Segmentation
• Rethinking Semantic Segmentation from a Sequence-to-Sequence Perspective with Transformers
• Dual Graph Convolutional Network for Semantic Segmentation
• Bi-directional Cross-Modality Feature Propagation with Separation-and-Aggregation Gate for RGB-D Semantic Segmentation
• Structured Consistency Loss for semi-supervised semantic segmentation
• A Comprehensive Review of Modern Object Segmentation Approaches
• Strip Pooling: Rethinking Spatial Pooling for Scene Parsing
• Global Aggregation then Local Distribution for Scene Parsing
• Vision Transformers: From Semantic Segmentation to Dense Prediction
• PointFlow: Flowing Semantics Through Points for Aerial Image Segmentation
• EfficientSeg: An Efficient Semantic Segmentation Network
• CaseNet: Content-Adaptive Scale Interaction Networks for Scene Parsing
• ARM: A Confidence-Based Adversarial Reweighting Module for Coarse Semantic Segmentation