Human Motion Modeling using DVGANs
arXiv:1804.10652
Abstract
We present a novel generative model for human motion modeling using Generative Adversarial Networks (GANs). We formulate the GAN discriminator using dense validation at each time-scale and perturb the discriminator input to make it translation invariant. Our model is capable of motion generation and completion. We show through our evaluations the resiliency to noise, generalization over actions, and generation of long diverse sequences. We evaluate our approach on Human 3.6M and CMU motion capture datasets using inception scores.
References in corpus (5)
- Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
- Unifying Visual-Semantic Embeddings with Multimodal Neural Language Models
- A Note on the Inception Score
- Boundary-Seeking Generative Adversarial Networks
- Language Generation with Recurrent Generative Adversarial Networks without Pre-training
Cited by in corpus (8)
- Diverse Human Motion Prediction Guided by Multi-Level Spatial-Temporal Anchors
- Human Motion Prediction via Learning Local Structure Representations and Temporal Dependencies
- DLow: Diversifying Latent Flows for Diverse Human Motion Prediction
- Learning Variations in Human Motion via Mix-and-Match Perturbation
- ArTIST: Autoregressive Trajectory Inpainting and Scoring for Tracking
- Generating Smooth Pose Sequences for Diverse Human Motion Prediction
- Deep Sequence Learning for Video Anticipation: From Discrete and Deterministic to Continuous and Stochastic
- A Comprehensive Multi-scale Approach for Speech and Dynamics Synchrony in Talking Head Generation