Multi-human Parsing with a Graph-based Generative Adversarial Model

Human parsing is an important task in human-centric image understanding computer vision and multimedia systems. However, most existing works on human mainly tackle the single-person scenario, which deviates from real-world applications where multiple persons are present simultaneously with interaction occlusion. To address such a challenging multi-human problem, we introduce novel model named MH-Parser, uses graph-based generative adversarial to challenges of close-person occlusion parsing. validate effectiveness new model, collect dataset Multi-Human Parsing (MHP), contains intensive person entanglement. Experiments MHP datasets demonstrate that proposed method effective addressing problem compared solutions literature.