The C3SR team won the 1st place for the CVPR Visual Understanding of Humans in Crowd Scene and the 2nd Look Into Person (LIP) Challenge for three Tacks: Track1 Look Into Person: Single Person Human Parsing Challenge(50462 images), Track2 Look Into Person: Multi-Person Human Parsing Challenge(38280 images) and Track5 Look Into Person: Fine-Grained Multi-Human Human Parsing Challenge(25403 images)