[Seminar] A Flow-Guided Mutual Attention Network for Video-Based Person Re-Identification