DyST：実世界動画における動的ニューラルシーン表現の学習
DyST: Towards Dynamic Neural Scene Representations on Real-World Videos
単眼の実世界動画から、シーン内容・視点ごとの動き・カメラ姿勢を潜在表現として分離するDynamic Scene Transformerを提案し、カメラとシーン内容を独立に制御した視点生成を可能にした。
著者: Maximilian Seitzer, Sjoerd van Steenkiste, Thomas Kipf, Klaus Greff, Mehdi S. M. Sajjadi
分類: cs.CV, cs.AI, cs.GR, cs.LG, cs.RO
原文アブストラクト
Visual understanding of the world goes beyond the semantics and flat structure of individual images. In this work, we aim to capture both the 3D structure and dynamics of real-world scenes from monocular real-world videos. Our Dynamic Scene Transformer (DyST) model leverages recent work in neural scene representation to learn a latent decomposition of monocular real-world videos into scene content, per-view scene dynamics, and camera pose. This separation is achieved through a novel co-training scheme on monocular videos and our new synthetic dataset DySO. DyST learns tangible latent representations for dynamic scenes that enable view generation with separate control over the camera and the content of the scene.
