English

Learning Video Dynamics with Predictive Differentiable Rendering

Computer Vision and Pattern Recognition 2026-06-30 v1 Artificial Intelligence

Abstract

How to accurately predict a high-fidelity future world? While the visual world is inherently continuous, existing deterministic video prediction models operate in discrete pixel space and are mainly optimized with pixel-wise mean squared error (MSE), which often leads to over-smoothed predictions and a lack of fine-grained visual details. To address these limitations, we propose Predictive Differentiable Rendering (PDR), a novel end-to-end video prediction paradigm that bridges the gap between discrete and continuous representations. Inspired by recent progress in 3D reconstruction with 3D Gaussian Splatting, we introduce PredGS, a lightweight and plug-and-play adapter based on 2D Gaussian representation, which could be seamlessly integrated with existing pixel space predictors, significantly improving spatial detail preservation with negligible computational overhead. Furthermore, we develop predgsplat, a CUDA-accelerated differentiable 2D Gaussian renderer supporting arbitrary channels. Each Gaussian is defined by 5 + C learnable parameters (position, scale, rotation, and C channel amplitudes) and achieves up to 10x faster rendering than the baseline. Optimized by a combined L1 and SSIM loss, PDR overcomes the inherent blurring tendencies of MSE Loss, significantly enhancing the prediction performance. Extensive experiments on diverse real-world benchmarks, including TaxiBJ, WeatherBench, KTH, and Human3.6M, demonstrate that PDR consistently surpasses existing methods, delivering superior detail preservation, visual fidelity, and predictive accuracy.

Cite

@article{arxiv.2606.31050,
  title  = {Learning Video Dynamics with Predictive Differentiable Rendering},
  author = {Yujin Tang and Tian Zhou and Xin Lin and Cheng Tan and Yifan Hu and Rong Jin and SouYoung Jin and Liang Sun},
  journal= {arXiv preprint arXiv:2606.31050},
  year   = {2026}
}

Comments

Accepted by ECCV 2026. 18 pages, 5 figures, 11 tables