Abstract:How to proper encode high-order object relation in the detection system without any external knowledge? How to leverage the information between co-occurrence and locations of objects for better reasoning? These questions are key challenges towards large-scale object detection system that aims to recognize thousands of objects entangled with complex spatial and semantic relationships nowadays. Distilling key relations that may affect object recognition is crucially important since treating each region separately leads to a big performance drop when facing heavy long-tail data distributions and plenty of confusing categories. Recent works try to encode relation by constructing graphs, e.g. using handcraft linguistic knowledge between classes or implicitly learning a fully-connected graph between regions. However, the handcraft linguistic knowledge cannot be individualized for each image due to the semantic gap between linguistic and visual context while the fully-connected graph is inefficient and noisy by incorporating redundant and distracted relations/edges from irrelevant objects and backgrounds. In this work, we introduce a Spatial-aware Graph Relation Network (SGRN) to adaptive discover and incorporate key semantic and spatial relationships for reasoning over each object. Our method considers the relative location layouts and interactions among which can be easily injected into any detection pipelines to boost the performance. Specifically, our SGRN integrates a graph learner module for learning a interpatable sparse graph structure to encode relevant contextual regions and a spatial graph reasoning module with learnable spatial Gaussian kernels to perform graph inference with spatial awareness. Extensive experiments verify the effectiveness of our method, e.g. achieving around 32% improvement on VG(3000 classes) and 28% on ADE in terms of mAP.

Neighbouring Relationship Exploration Based on Graph Convolutional Network for Object Detection

Visual relationship detection with a deep convolutional relationship network

Object Detection by Crossing Relational Reasoning Based on Graph Neural Network

Mining graph-based dynamic relationships for object detection

GAR: Graph Assisted Reasoning for Object Detection

DIGCN: A Dynamic Interaction Graph Convolutional Network Based on Learnable Proposals for Object Detection

One-stage Object Detection with Graph Convolutional Networks

Residual Joint Attention Network with Graph Structure Inference for Object Detection.

Multiview Diffusion-Based Affinity Graph Learning with Good Neighbourhoods for Salient Object Detection

Spatial-Aware Graph Relation Network for Large-Scale Object Detection.

C-FCN: Corners-based Fully Convolutional Network for Visual Object Detection

DAGCN: Dynamic and Adaptive Graph Convolutional Network for Salient Object Detection.

Relation Graph Network for 3D Object Detection in Point Clouds

A Novel Dynamic Graph Evolution Network for Salient Object Detection

Dynamically Connected Graph Representation for Object Detection.

Cascaded Convolutional Neural Networks for Object Detection.

Relation Networks for Object Detection

A Double-Head Global Reasoning Network for Object Detection of Remote Sensing Images

Object Detection Via Inner-Inter Relational Reasoning Network.

Salient Object Detection Based on Unified Graph Neural Network Joint Learning

Semantic-Context Graph Network for Point-based 3D Object Detection