Phase-functioned neural networks for character control

By Daniel Holden, Taku Komura, Jun Saito. 2017

668 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Control and Systems Engineering

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Control and Systems Engineering

Citations in the corpus

Phase-Parametric Policies for Reinforcement Learning in Cyclic Environments (2018, methodology): “Recurrent Neural Networks (RNNs) have the ability to retain information about previously seen inputs.”
Mode-adaptive neural networks for quadruped motion control (2018, methodology): “Instead of training the character to walk over different terrains as done in [Holden et al. 2017], we simply apply a CCD full-body inverse kinematics to the character movements.”
Language2Pose: Natural Language Grounded Pose Forecasting (2019, background): “Pose forecasting for locomotion is a more commonly researched topic, where models decide where and when to run/walk based on low-level control parameters such as trajectory and terrain [13].”
Generating Diverse and Natural 3D Human Motions from Text (2022, methodology): “We apply the same pose processing steps as in [13].”
AI Choreographer: Music Conditioned 3D Dance Generation with AIST++ (2021, background): “Phase-functioned neural networks and it’s variations [94, 33, 73, 74] address this issue via conditioning the network weights on phase, however, they do not scale well to represent a wide variety of motion.”
AMP (2021, methodology): “Data-driven methods using generative models, such as Gaussian processes [Levine et al. 2012; Ye and Liu 2010] and neural networks [Holden et al. 2017; Ling et al. 2020; Zhang et al. 2018], have also been applied to synthesize motions online.”
Generating Human Motion from Textual Descriptions with Discrete Representations (2023, methodology): “[32] proposes phase functioned neural network to perform the control task.”
Hierarchical Style-based Networks for Motion Synthesis (2020, methodology): “ataset. Meanwhile, it is hard for these approaches to generalize to unseen distribution of data. To adapt and generalize to large-scale data, deep neural networks are utilized in several recent works [11,10,41]. For example, Holden et al. [11] introduced to synthesize the motion sequence using deep ”
Learning human behaviors from motion capture by adversarial imitation (2017, background): “Other recent work successfully makes use of a specialized neural network architecture for motion capture sequence modeling to enable robust humanoid behavior [22].”
HuMoR: 3D Human Motion Model for Robust Pose Estimation (2021, background): “imation [12, 48, 79, 50, 54, 37, 86] including recent recurrent and autoregressive models [32, 28, 35, 101, 52].”
Robust motion in-betweening (2020, background): “Online control from a gamepad has also been tackled with phase-aware [Holden et al. 2017], mode-aware [Zhang et al. 2018] and action-aware [Starke et al. 2019] neural networks that can automatically choose a mixture of network weights at run-time to disambiguate possible motions.”
Action-Conditioned 3D Human Motion Synthesis with Transformer VAE (2021, background): “Most relevant here are the phase-functioned neural networks [26] and neural state machines [57].”
TM2T: Stochastic and Tokenized Modeling for the Reciprocal Generation of 3D Human Motions and Texts (2022, methodology): “Meanwhile, [19,18,42] attempts to model human motions according to instant control signals such as velocity and directional readouts.”
Hierarchical visuomotor control of humanoids (2018, background): “While the resulting time-indexed policies are somewhat less general as a result, time-indexing or phase-variables are common in the animation literature and also employed in kinematic control of characters (Holden et al., 2017).”
Policies Modulating Trajectory Generators (2018, methodology): “In computer graphics, special architectures, such as Phase-Functioned Neural Networks [11] and Mode-Adaptive Neural Networks [12], have been proposed to synthesize locomotion controllers from motion capture data.”
Character controllers using motion VAEs (2020, methodology): “Autoregressive DNN models can produce high-quality human variable-terrain locomotion [Holden et al. 2017], quadruped variable-terrain locomotion [Zhang et al. 2018], and environment aware human locomotion [Starke et al. 2019].”