首页> 外国专利> METHODS AND SYSTEMS FOR 3D CHARACTER POSE PREDICTION BASED ON PHASE-FUNCTIONED NEURAL NETWORKS

METHODS AND SYSTEMS FOR 3D CHARACTER POSE PREDICTION BASED ON PHASE-FUNCTIONED NEURAL NETWORKS

机译:基于相位功能神经网络的3D字符姿态预测方法和系统

摘要

Disclosed is a method of 3D character pose prediction based on phase-functioned neural networks, which generally relates to 3D pose prediction and AI-driven animation. In the present system, a character control neural network can be used to automatically produce animation. Using the present system, a user can set up a character by defining the path, velocity and/or body shape of the character. After setting up the character, the network can produce skeletal position and rotation information for each frame. The network output can be adjusted with inverse kinematics (IK). If a forecast result goes beyond reasonable expectations, the system can adjust this bone's position and recalculate the angles at other joints.
机译:公开了一种基于相位功能神经网络的3D字符姿态预测方法,其通常涉及3D姿态预测和AI驱动的动画。 在本系统中,可以使用一个字符控制神经网络来自动产生动画。 使用本系统,用户可以通过定义字符的路径,速度和/或身体形状来设置字符。 在设置字符之后,网络可以为每个帧产生骨架位置和旋转信息。 可以使用逆运动学(IK)调整网络输出。 如果预测结果超出了合理的期望,则系统可以调节该骨骼的位置并在其他关节处重新计算角度。

著录项

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号