Robotic deformable-object manipulation is a challenge in the robotic industry because deformable objects have complicated and various object states. Predicting those object states and updating manipulation planning is time-consuming and computationally expensive. In this paper, we propose learning known configurations of garments to allow a robot to recognise garment states and choose a pre-designed manipulation plan for garment flattening.