简约风科技场景手部动作数据插画素材

Draw a schematic diagram of a dexterous hand grasping pipeline using Mocap, RL, and ECN, in academic conference style (ICRA/RSS/CoRL). - Left: "MoCap MANO Hand Data" → arrow → "Retarget to DexHand" → "DexHand Trajectory".- DexHand Trajectory branches into two: 1. Arrow → "ECN (Expected Contact Network)" → outputs predicted contact sequence [s_t^i, df_t^i]. 2. Arrow → "RL Agent" (policy network), receives state input.- ECN output connects as auxiliary input / reward to RL: arrow labeled "contact prediction / reward guidance".- RL Agent outputs optimized hand motion → arrow → "DexHand Motion Optimized for Grasping".- Style: professional, minimalistic, academic figure, clear boxes and arrows, readable labels, color-code: - Mocap input: blue - DexHand: green - ECN: pink - RL agent: light green- Emphasize sequence processing and auxiliary feedback from ECN to RL.