RBFNN Pose Interpolation with Angular Distance Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for animating digital characters in virtual reality and augmented reality face challenges in accurately interpolating poses due to large data requirements, resource constraints, and periodic errors in radial basis function neural networks (RBFNNs), leading to unstable and unpredictable animations.
Innovation Solution
The use of radial basis function neural networks (RBFNNs) with separation of linear and angular parameters, combined with weight application and quaternion vectors to represent angular values, addresses the challenges by enabling more accurate interpolation and reducing resource usage, and stabilizing the animation process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If traditional interpolation methods are used to animate digital character poses, then resource requirements and data storage needs increase, but computing speed and efficiency deteriorate
Solution Approach 1:
The patent segments the pose space into discrete training poses that sample the overall pose volume, rather than storing or processing all possible poses. The RBFNN interpolation engine then segments the computation by calculating distances only to relevant training poses and combining them weighted by distance, reducing both data requirements and computing speed requirements.
2Productivity
If RBFNN is used for pose interpolation, then computing speed improves, but periodic errors and animation stability deteriorate
Solution Approach 1:
The patent implements feedback through the distance calculation and weight combination mechanism. The system continuously calculates the distance between the current pose and training poses, determines appropriate weights based on these distances, and combines the training pose data accordingly. This feedback loop ensures that the interpolation adapts to the pose space geometry and maintains animation stability by relying on verified distance relationships rather than direct neural network outputs.
3Device complexity
If linear and angular parameters are mixed in interpolation, then device complexity reduces, but manufacturing precision and interpolation accuracy deteriorate
Solution Approach 1:
The patent segments the parameter space by separating linear and angular parameters into distinct distance calculation processes. Linear parameters use Euclidean distance metrics while angular parameters use appropriate angular distance metrics. This segmentation allows each parameter type to be processed with its most suitable mathematical treatment, improving interpolation accuracy without significantly increasing overall system complexity.
Data Source
AI summary
Systems and methods are provided for interpolation of disparate inputs. A radial basis function neural network (RBFNN) may be used to interpolate the pose of a digital character. Input parameters to the RBFNN may be separated by data type (e.g. angular vs. linear) and manipulated within the RBFNN by distance functions specific to the data type (e.g. use an angular distance function for the angular input data). A weight may be applied to each distance to compensate for input data representing different variables (e.g. clavicle vs. shoulder). The output parameters of the RBFNN may be a set of independent values, which may be combined into combination values (e.g. representing x, y, z, w angular value in SO(3) space).