Pose-Aware Placement of Objects with Semantic Labels - Brandname-based Affordance Prediction and Cooperative Dual-Arm Active Manipulation.