Inductive Graph Neural Networks for Moving Object Segmentation

TL;DR

This paper introduces GraphIMOS, an inductive graph neural network approach for moving object segmentation that generalizes to new data, outperforming previous methods and enabling real-world deployment.

Contribution

The paper presents a novel inductive GNN-based algorithm for MOS that generalizes to unseen data, unlike prior transductive methods.

Findings

GraphIMOS outperforms previous inductive methods.

The approach enables real-world application of graph-based MOS.

It effectively handles new data frames during deployment.

Abstract

Moving Object Segmentation (MOS) is a challenging problem in computer vision, particularly in scenarios with dynamic backgrounds, abrupt lighting changes, shadows, camouflage, and moving cameras. While graph-based methods have shown promising results in MOS, they have mainly relied on transductive learning which assumes access to the entire training and testing data for evaluation. However, this assumption is not realistic in real-world applications where the system needs to handle new data during deployment. In this paper, we propose a novel Graph Inductive Moving Object Segmentation (GraphIMOS) algorithm based on a Graph Neural Network (GNN) architecture. Our approach builds a generic model capable of performing prediction on newly added data frames using the already trained model. GraphIMOS outperforms previous inductive learning methods and is more generic than previous transductive techniques. Our proposed algorithm enables the deployment of graph-based MOS models in real-world applications.