Language2Pose: Natural Language Grounded Pose Forecasting

  title={Language2Pose: Natural Language Grounded Pose Forecasting},
  author={Chaitanya Ahuja and Louis-Philippe Morency},
  journal={2019 International Conference on 3D Vision (3DV)},
  • Chaitanya Ahuja, Louis-Philippe Morency
  • Published 2019
  • Computer Science
  • 2019 International Conference on 3D Vision (3DV)
  • Generating animations from natural language sentences finds its applications in a a number of domains such as movie script visualization, virtual human animation and, robot motion planning. [...] Key Method This joint embedding space is learned end-to-end using a curriculum learning approach which emphasizes shorter and easier sequences first before moving to longer and harder ones. We evaluate our proposed model on a publicly available corpus of 3D pose data and human-annotated sentences. Both objective…Expand Abstract

    Figures, Tables, and Topics from this paper.

    Action2Motion: Conditioned Generation of 3D Human Motions
    Coalescing Narrative and Dialogue for Grounded Pose Forecasting
    Dynamic Future Net: Diversified Human Motion Generation
    Sequence-to-Sequence Predictive Model: From Prosody To Communicative Gestures
    3D Human Motion Estimation via Motion Compression and Refinement


    Publications referenced by this paper.
    A deep learning approach for generalized speech animation
    • 112
    • PDF
    Jointly Modeling Embedding and Translation to Bridge Video and Language
    • 339
    • PDF
    The KIT Motion-Language Dataset
    • 14
    • PDF
    Text2Action: Generative Adversarial Synthesis from Language to Action
    • 19
    • PDF
    Paired Recurrent Autoencoders for Bidirectional Translation Between Robot Actions and Linguistic Descriptions
    • 13
    Generating Animations from Screenplays
    • 3
    • PDF
    Action-Agnostic Human Pose Forecasting
    • 28
    • PDF
    Learning Human Motion Models for Long-Term Predictions
    • 82
    • PDF