Foreground Image Segmentation with FgSegNet

A multi-scale CNN with an autoencoder-like architecture

A diagram of FgSegNet’s architecture; it uses three CNN models and a single TCNN. [1]

Introduction

A tough challenge that remains to be solved robustly is foreground image segmentation (or background subtraction in a different perspective). The task may sound trivial: create a binary mask where only the pixels of a moving/important object are…