首页|基于多特征融合的虚拟数字人物动作捕捉技术

基于多特征融合的虚拟数字人物动作捕捉技术

扫码查看
为准确检测虚拟数字人物动作姿态特征点,降低动作捕捉数据的计算误差,提出基于多特征融合的虚拟数字人物动作捕捉技术.采用动作最小化函数解算人物动作姿态,引入尺度空间向量检测人体动作姿态特征点,结合人体运动模型跟踪人物动作.以此为依据,利用多融合特征算法计算动作捕捉特征量,并匹配计算结果与动作轨迹模板,以此完成人物动作捕捉.实验结果表明,所提方法测得的动作捕捉数据与实际数据的误差较小,捕捉精度较高.
Motion Capture Technology of Virtual Digital Characters Based on Multi-Feature Fusion
To accurately detect the dynamic pose feature points of virtual digital characters and reduce the computational errors of motion capture data,a virtual digital character motion capture technology based on multi-feature fusion is proposed.Using the action minimization function to calculate the posture of a person's actions,introducing scale space vectors to detect the dynamic posture feature points of the human body,and combining the human motion model to track the person's actions.Based on this,a multi fusion feature algorithm is used to calculate the feature quantity of action capture,and the calculation results are matched with the action trajectory template to complete the capture of character actions.The experimental results show that the error between the action capture data measured by the proposed method and the actual data is small,and the capture accuracy is high.

multi-feature fusionvirtual digital charactersmotion capturedynamic capture accuracy

吴建美

展开 >

福州软件职业技术学院 游戏产业学院,福建 福州 350005

多特征融合 虚拟数字人物 动作捕捉 动捕精度

2024

黑龙江工业学院学报(综合版)
鸡西大学

黑龙江工业学院学报(综合版)

影响因子:0.211
ISSN:1672-6758
年,卷(期):2024.24(1)
  • 17